CIT 212

NOUN · System Analysis and Design

CIT 212 Study Guide

Rewritten and checked version of NOUN CIT 212: 3 modules, 12 units, worked examples, common mistakes, the course’s errata, and practice questions with the answers folded away.

Computer-based exam: MCQ and fill-in-the-blank · answers hidden until opened

This page holds the whole course in one file. Every unit keeps the chapter’s six parts in the same order: what you must know, the key ideas, the code and worked examples, the common mistakes, the errata, and practice questions. The chapter wording is copied verbatim; only this page’s chrome (masthead, contents, colours, code colouring) is written for the web.

Reading this page

  • Try before you look. Every answer set is folded into a Answers box; open it only after you have written your own answer.
  • Links land on the unit. A link such as #m2u3 (the Anki deck uses these) scrolls straight to that unit and leaves it highlighted for a few seconds.
  • Code scrolls sideways. Long code lines stay on one line inside their own box; the page itself never scrolls sideways. Copy takes the exact source, without the colouring.
  • Errata belong to the chapter. The course’s version is kept in Key ideas, because that is what the exam marks; the correction is in that unit’s Errata.

Modules at a glance

System Analysis and Design has 3 modules and 12 units. Each module below has its own units, its worked examples and its external reading list; the ids in the contents are the anchors the Anki deck links to.

ModuleTitleUnits
Module 1System Analysis and Requirements4
Module 2System Design (Input, Output, Files)4
Module 3Development, Quality and Implementation4

Read the unit → work the examples by hand → answer the practice questions without opening Answers → check → re-do the wrong ones.

Module 1 of 3 · 4 units · PDF pp.19–111

Module 1 — System Analysis and Requirements

Page convention. The course pages carry printed numbers, but "p.N" in this guide always means the PDF page N of CIT212.pdf (the =====PAGE N===== marker in the extracted text). Printed page = PDF page − 18. Module 1 is PDF pp.19–111: Unit 1 pp.19–38, Unit 2 pp.39–55, Unit 3 pp.56–76, Unit 4 pp.77–111. Figure numbers are the course's own and are not reliable in Unit 4 (the figures are labelled 3.x but the text calls them 4.x; see the Unit 4 Errata).

Likely to be examined. The exam is computer-based, MCQ plus fill-in-the-blank. The course does not say which topics are examined, so use its own question sets as the best signal: the unit objectives (pp.20, 40, 57, 78), Self-Assessment Exercise 1 (p.28, four questions) and 2 (p.33, four questions) in Unit 1, the Self-Assessment Exercise in Unit 2 (pp.42–43, five questions) and Unit 3 (p.62, four questions), and the Tutor-Marked Assignments (Unit 1 p.38, four questions; Unit 2 p.54, four; Unit 3 p.76, three; Unit 4 p.111, four). Unit 4 has no self-assessment exercise. Every list in this module (characteristics, elements, types, rules, steps) is the kind of thing a fill-in item asks for, so learn the counts as well as the items.

Jump to unit1234

Module 1 · Unit 1 of 4

Unit 1 — Overview of System Analysis and Design

What you must know

  • Define system, systems study, systems analysis and systems approach, and say how the last two differ (p.21).
  • State the five characteristics common to all systems (organisation, interaction, interdependence, integration, central objective) (pp.21–22).
  • Name the four elements of systems analysis (output, inputs, files, processes) and the four essentials of inputs (pp.22–23).
  • Classify systems: physical/abstract, open/closed, deterministic/probabilistic, man-made information systems; and the four computer-based information systems TPS, MIS, DSS, OAS (pp.24–28).
  • List the six SDLC activities and what each one does, including the three parts of the preliminary investigation and the four dimensions of evaluation (pp.28–33).
  • Explain the software crisis from the programmer's and the user's point of view (pp.33–35).
  • Describe who a systems analyst is, what the analyst does (a)–(g), and the seven attributes of an effective analyst (pp.35–37).

Key ideas

Intuition. A system is a set of parts that only make sense together: each part does its own share, and the whole exists to reach one goal. Systems analysis works out what the system must do; systems design works out how (p.20).

Systems analysis and design (p.20). "The process of examining a business situation with the intent of improving it through better procedures and methods." It has two components, systems analysis and systems design.

  • Systems design is "the process of planning a new system or replace or complement an existing system".
  • Systems analysis is "the process of gathering and interpreting facts, diagnosing problems and using the information to recommend improvement to the system".
  • "Analysis specified what the system should do. Design states how to accomplish the objectives."

System (pp.20–21). From the Greek systema, "an organised relationship among the functioning units or components". Most definitions agree: "a system is an orderly grouping of interdependent components linked together, according to a plan, to achieve a specific goal." A component may be a physical part (engine, wheels), a managerial step (planning, organising, controlling) or a subsystem in a multi-level structure. A business system combines policies, personnel, equipment and computer facilities to coordinate the activities of the organisation (p.20).

Term Course definition (p.21)
Systems study "a study of the operations of a set of connected elements and of the inter-connections between these elements"; no part can be ignored without finding out its effect on the whole.
Systems approach "a set of procedure for solving a particular problem": applies scientific methods to observe, clarify, identify and solve a problem, with care for the inter-relatedness of elements.
Systems analysis "a management technique which helps us in designing a new system or improving an existing system".

Five characteristics of a system (pp.21–22).

Characteristic Meaning in the course
a) Organisation Structure and order; the arrangement of components that helps achieve objectives (e.g. president down to blue-collar workers; a computer built around input device, CPU, output device and storage units).
b) Interaction How each component functions with the others (purchasing with production, advertising with sales, payroll with personnel; CPU with memory).
c) Interdependence Components depend on one another; coordinated and linked in a planned way to reach an objective.
d) Integration How the system is tied together; more than sharing a physical part or location; parts work together though each does a unique function; produces a better result than components working independently.
e) Central objective "The last characteristic"; objectives may be real or stated; an organisation may state one objective and operate toward another; users must know the central objective in advance.

Four elements of systems analysis (pp.22–23).

  1. Output. Decide the objectives or goals first ("if we do not know where we want to go, we will not know when we have reached there"). The user department defines them as needs; these become the outputs the analyst keeps in mind.
  2. Inputs. Once the output is known, the inputs follow. If required information is costly to get, prepare a cost-benefit analysis to convince management. Essential elements of inputs: (i) accuracy, (ii) timeliness, (iii) proper format, (iv) economy (data produced at the least cost).
  3. Files. Store data (historical or generated within the system), as isolated facts or in large volumes.
  4. Processes. How inputs and files are converted into outputs: a set of logical steps given to the computer as "programs".

Types of systems (pp.23–25). Four common classifications:

Pair First Second
Physical or abstract Physical: tangible entities, static or dynamic in operation Abstract: conceptual, non-physical (formulas, relationships among sets of variables, models)
Open or closed Open: continually interacts with its environment, receives inputs from and delivers outputs to the outside; an information system is open because it must adapt to changing user demands Closed: isolated from environmental influences; "in reality completely closed systems are rare"
Deterministic or probabilistic Deterministic: all events perfectly predictable; from the state at one time the next state is predicted; example: a numerically controlled machine tool Probabilistic: events cannot be perfectly predicted; example: a warehouse and its contents
Man-made information systems An information system is "a set of devices, procedures and operating systems designed around user-based criteria to produce information and communicate it to the user for planning, control and performance" Information "reduces uncertainty about a state or event" (example: calm wind reduces uncertainty about an enjoyable boat trip)

Major information systems in a business (pp.24–25). A business has several, each for a specific purpose:

  • Formal information system: based on the organisation chart (boxes and straight lines); concerned with authority, communication and work flow.
  • Informal information system: employee-based, meets personnel and vocational needs, helps solve work-related problems, funnels information upward through indirect channels.
  • Computer-based information systems (CBIS): four types (p.25).

The four computer-based information systems (pp.25–28).

System Course description Characteristics (Table 1.1, p.28)
TPS Transaction Processing System "a computer-based system that captures, classifies, stores, maintains, updates and retrieves transaction data for record keeping and for input to other types of CBIS" (p.26). A transaction is "any event or activity that affects the whole organization" (orders, billing, hiring, depositing cheques). Gives speed and accuracy and follows routines without variance. Substitutes computer-based processing for manual procedures; well-structured routine processes; record-keeping.
MIS Management Information System Provides "all levels of management with information essential to the running of smooth business", which must be "relevant, timely, accurate, complete and concise as is economically feasible" (p.27). Input to the managerial decision process; supports well-structured decision situations; typical information requirements can be anticipated.
DSS Decision Support System Offers information that "may not be predictable", that managers may need "only once"; no regularly scheduled reports; responds to a wide range of requests; "supports, but does not replace, judgement of manager" (p.27). Information for managers who make judgements about particular situations; supports decision makers in situations that are not well-structured.
OAS Office Automation System Among the newest and most rapidly expanding; aims to raise efficiency and productivity of office workers; word processing, electronic mail; "a multi-function, integrated computer-based system that allows many office activities to be performed in an electronic mode" (pp.27–28). Multi-function, integrated, office activities in an electronic mode.

A decision is unstructured "if there are no clear procedures for making the decision and if not all the factors to be considered in the decision can be readily identified in advance" (p.27). Figure 1.2 (p.26), the hierarchical view of CBIS: Top management → DSS; Middle management → MIS; First-line management → TPS; Clerical personnel → OAS. Figure 1.1 shows the CBIS in its organisational context (planning, controlling and decision making; sales and receipts; clerical tasks).

System Development Life Cycle (SDLC) (pp.28–33). System development is the two major steps of analysis and design, and "starts when management or sometimes system development personnel feel that a new system or an improvement in the existing system is required" (pp.28–29). The SDLC is "the set of activities that analysts, designers and users carry out to develop and implement an information system". Six activities (p.29):

  1. Preliminary investigation
  2. Determination of system requirements
  3. Design of system
  4. Development of software
  5. Systems testing
  6. Implementation, evaluation and maintenance
Activity What the course says
1. Preliminary investigation (pp.29–30) Begins when someone makes a request. Three parts: (i) request clarification (requests are often not clearly defined, so examine and clarify them), (ii) feasibility study, (iii) request approval (feasible and desirable projects are scheduled; costs, priority, completion time and personnel requirements are estimated).
2. Determination of system requirements (pp.30–31) "At the heart of systems analysis". Key questions: what is being done, how, how frequently, how great the volume of transactions or decisions, how well is the task performed, does a problem exist, how serious and what cause. Tools: data flow diagrams (DFD), interviews, on-site observations, questionnaires; also study of manuals and reports. Result: the analyst "has a firm understanding of what is to be done".
3. Design of system (p.31) Describes how the system will meet the requirements. Called logical design "in contrast to the process of developing program software, which is referred to as physical design". Starts from the reports and outputs, then the data on each; designers select file structures and storage devices; specifications go to the programmers.
4. Development of software (pp.31–32) Install purchased software or develop custom programs; the choice depends on cost, time available and availability of programmers. Programmers document programs; documentation is essential for testing and maintenance.
5. Systems testing (p.32) The system is used experimentally "to ensure that the software does not fail"; special test data are input and results examined; a limited number of users may try it to reveal unforeseen use.
6. Implementation, evaluation and maintenance (pp.32–33) See below.

Feasibility study in the preliminary investigation: three aspects (pp.29–30).

Feasibility Question the course asks
Technical Can the work be done with current equipment, existing software technology and available personnel? If new technology is needed, what is the likelihood it can be developed?
Economic Are there sufficient benefits to make the costs acceptable? Or are the costs of not creating the system so great that it is advisable to undertake the project?
Operational Will the system be used if developed and implemented? Will user resistance undermine the benefits?

It is carried out by a small group familiar with information-systems techniques, who understand the business parts involved and are skilled in analysis and design (p.30).

Implementation (p.32). Systems personnel check out and put new equipment into use, train users, install the new application and construct the data files needed. Changeover choices named: test in only one area of the firm with one or two persons; run old and new systems in parallel to compare results; stop the old system one day and start the new one the next day.

Evaluation: four dimensions (pp.32–33).

Dimension Assesses
(i) Operational evaluation How the system functions: ease of use, response time, overall reliability, level of utilisation.
(ii) Organizational impact Benefits to the organisation: financial concerns, operational efficiency, competitive impact.
(iii) User manager assessment Attitudes of senior and user managers and of end-users.
(iv) Development performance The development process against yardsticks: overall development time and effort, conformance to budgets and standards, other project-management criteria.

Maintenance (p.33). Necessary "to eliminate errors in the working system during its working life and to tune the system to any variations in its working environment". Planners must plan resource availability for it.

Software crisis (pp.33–35). Hardware became cheaper and more accessible, creating a matching demand for software. "The traditional intuitive and ad-hoc approach fails miserably when the quantities of data involved in information systems exceeds say, 10 MB", the figure at which systems pass from relatively simple to significant complexity (p.33). This led to the phrase "software crisis" and the search for structured methods (structured systems development methodology; software engineering).

From the programmer's point of view (p.34, eight items) From the user's point of view (pp.34–35, twelve items)
Compatibility; portability; documentation; coordination of the work of different people in a team; run-time errors not detected in the sample run; piracy; customers expanding specifications after design and implementation; proper maintenance Choosing software from the total market; ensuring compatibility with hardware; customised software not meeting total requirements; viruses; bugs found only after considerable data entry; software that runs only on a specific operating system; compatibility of floppy diskettes (size and density); learning all facilities when manuals give selective information; files that expand memory use and cause disk-management problems; software memory requirement above the existing requirement or availability; different versions of software (user and operating system); security of protected data

The systems analyst (pp.35–37).

  • Who (p.35): "a person who conducts a study, identifies activities and objectives and determines a procedure to achieve the objectives"; designs and implements systems to suit organisational needs; "plays a major role in seeing business benefits from computer technology"; coordinates the efforts of different people.
  • What the analyst does (pp.35–36), (a)–(g): (a) problem definition, "the first and perhaps most difficult task"; (b) consults managers, users and other data-processing professionals, using various data-gathering methods; (c) analyses the data and thinks of a plan, pulling together other people's ideas and refining them; (d) coordinates the development of solutions and evaluates the merit of each before recommending one; (e) is a planner; (f) once the plan is accepted, designs it ("time consuming, complex and precise"); (g) coordinates the testing and helps decide whether the system meets standards set in the planning phase.
  • Seven attributes (pp.36–37): (a) knowledge of people; (b) knowledge of business functions (accounting and marketing in particular); (c) knowledge of data-processing principles (potential and limitations of computers); (d) ability to communicate; (e) flexibility; (f) an analytical mind; (g) well educated with sharp mind.

Code and worked examples

No program code in this unit; the worked examples are classifications and a trace of the SDLC.

Example 1 — Deterministic or probabilistic? Open or closed? (Course example p.24; classification table is mine). Apply the course's two tests. Deterministic = from the present state the next state can be predicted exactly. Open = it takes inputs from and delivers outputs to its environment.

System Deterministic / probabilistic Open / closed Reason
Numerically controlled machine tool (course p.24) Deterministic — Course names it as the deterministic example.
Warehouse and its contents (course p.24) Probabilistic Open Course names it as the probabilistic example: stock movements cannot be perfectly predicted.
A computer-based information system — Open Course p.24: it "must adapt to the changing demands of the user".
An isolated, self-contained model — Closed Course p.24: "completely closed systems are rare".

Example 2 — Which CBIS? (Course Table 1.1, p.28 and Figure 1.2, p.26; scenarios are my illustrations).

Scenario System Why
A clerk records fee payments and prints receipts TPS Routine, well-structured record keeping; clerical/first-line level.
Each month the bursar receives a standard list of fee defaulters MIS Regular report to management; requirements can be anticipated.
The registrar asks one-off "what if the intake rises 10 %?" questions DSS Unstructured decision; judgement stays with the manager.
Staff write and e-mail memos from word-processing software OAS Office activities performed in an electronic mode.

Check against Figure 1.2: first-line management → TPS, middle management → MIS, top management → DSS, clerical personnel → OAS.

Example 3 — Trace one request through the SDLC (course's activities, pp.28–33; the "student records" request is my illustration).

  1. Preliminary investigation. A registrar asks for computerised student records. Request clarification: what exactly is wanted? Feasibility: technical (equipment and staff?), economic (benefits against cost, and the cost of doing nothing?), operational (will staff use it?). Request approval: management schedules it.
  2. Determination of requirements. Interviews, observation, questionnaires, DFDs; ask what is done, how often, how many transactions, where the problem is.
  3. Design. Fix the reports and outputs first, then data items, files and procedures (logical design); specifications pass to programmers.
  4. Development of software. Buy a package or write custom programs; document them.
  5. Testing. Run special test data; let a few users try it.
  6. Implementation, evaluation, maintenance. Train users, install, build the files; changeover by pilot, parallel or direct; evaluate on the four dimensions; correct errors and tune it.

Example 4 — Match the statement to the evaluation dimension (course p.32–33; statements are mine).

  • "Response time is 3 seconds and the system rarely fails" → operational evaluation.
  • "Costs fell and the firm won tenders it used to lose" → organizational impact.
  • "The head of department says staff like the new screens" → user manager assessment.
  • "The project finished two months late and over budget" → development performance.

Common mistakes

  • Analysis vs design vs approach. Analysis = what (gathering and interpreting facts); design = how. "Systems approach" is a set of procedures (scientific method) for solving a problem; "systems analysis" is a management technique for designing or improving a system (p.21). The SAE 1(i) question asks exactly this.
  • "Logical design" in this course. The course calls the design stage logical design and software development physical design (p.31). An MCQ copying the course will mark that pairing correct; do not swap it for a textbook definition.
  • Elements of inputs vs elements of analysis. Output, inputs, files, processes are the four elements of systems analysis (p.22); accuracy, timeliness, proper format, economy are the four elements of inputs (p.23). Both lists have four items, which makes a swap easy.
  • Three feasibility aspects here, seven in Unit 3. Unit 1 (and Unit 2) name technical, economic, operational. Unit 3 names seven. Do not carry the three into a question that says "seven".
  • TPS/MIS/DSS/OAS by level. First-line = TPS, middle = MIS, top = DSS, clerical = OAS (Fig 1.2). A DSS does not replace the manager's judgement; an MIS serves well-structured decisions.
  • Open vs closed. An information system is open. Completely closed systems are rare.

Errata

Page Course says Correct Note
p.19 Unit heading "UNIT 1 SYSTEM ANALYSIS"; contents 3.2.1 "Preliminary of System Requirements" Module list (p.19) calls it "Overview of System Analysis and Design"; section 3.2.1 is "Preliminary Investigation" (p.29) Heading and contents differ from the body; this guide uses the module-list title.
p.20 "Systems Analysts and Systems Design" Systems Analysis and Systems Design Two components named in the next sentences.
p.26 "depending cheques" depositing cheques Typo in the list of common transactions.
p.30 "section 2.5 of unit 2" Preliminary investigation is section 3.4 of Unit 2 Unit 2 has no section 2.5.
p.32 "case of use" ease of use Operational evaluation.

Practice questions

  1. (SAE 1 i) "Systems approach" is a set of procedures for solving a problem by scientific method, while "systems analysis" is ______.
  2. (SAE 1 ii) The four basic elements of systems analysis are output, ______, files and processes.
  3. (SAE 1 iii) Which of these is the best description of a computer-based information system? A. Any system that has a computer in the building B. A system that depends mainly on the computer for handling business applications C. The organisation chart of a company D. A set of employee friendships

  4. (SAE 1 iv) A decision is unstructured when there are no clear ______ for making it and not all factors can be identified in advance.

  5. (SAE 2 i) How many activities make up the SDLC in this course? A. 4 B. 5 C. 6 D. 7
  6. (SAE 2 ii) The three feasibilities studied in the preliminary investigation are technical, economic and ______.
  7. (SAE 2 iii) "Ease of use, response time, overall reliability and level of utilisation" are the areas of which evaluation? A. Organizational impact B. User manager assessment C. Development performance D. Operational evaluation
  8. (SAE 2 iv / maintenance) Maintenance is necessary to ______ errors in the working system and to ______ the system to variations in its environment.
  9. Which system type is the course's example of a probabilistic system? A. Numerically controlled machine tool B. A warehouse and its contents C. A formula D. A calculator
  10. (TMA 2) A system's memory requirement exceeds the available memory. Does the course count this as a software problem? A. No, it is only a hardware fault B. Yes, it is listed among the software crisis problems C. No, it is an operational evaluation D. Yes, it is a TPS
Answersshow
  1. A management technique that helps in designing a new system or improving an existing one (p.21).
  2. Inputs (p.22).
  3. B; CBIS depends mainly on the computer for business applications and has four types: TPS, MIS, DSS, OAS (p.25).
  4. Procedures (p.27).
  5. C; six: preliminary investigation, determination of requirements, design, development of software, testing, implementation/evaluation/maintenance (p.29).
  6. Operational (p.29).
  7. D; operational evaluation (p.32).
  8. Eliminate; tune (p.33).
  9. B; the warehouse and its contents (p.24).
  10. B; the course lists "software crisis develops when system memory requirement of software is more than the existing requirements and/or availability" among the user's problems (p.35).

Module 1 · Unit 2 of 4

Unit 2 — Project Selection

What you must know

  • Give the five reasons for systems projects and the sub-reasons under each (pp.40–41).
  • State the four sources of project requests and why a request from senior executives weighs more (pp.41–42).
  • Describe the three review committees and what a project request form contains (pp.43–44).
  • State the purpose and five objectives of the preliminary investigation and its three data-gathering methods (pp.44–46).
  • Test a project on operational, technical and economic feasibility, and say what happens to infeasible projects (pp.46–48).
  • Define the seven classic system problems (reliability, validity, accuracy, economy, timeliness, capacity, throughput) and tell capacity from throughput (pp.48–52).
  • Evaluate a problem, know the 75 % / 25 % rule, and list sources of problems/opportunities and what to scan (pp.52–53).

Key ideas

Intuition. Analysts do not pick projects; requests come in, a committee chooses, and a short preliminary investigation decides whether a request is worth a full study (p.39).

Why system projects? Five reasons (pp.40–41).

Reason Sub-reasons (as the course lists them)
(a) Capability Information systems add capability in three ways (SAE ii): (i) improved processing speed; (ii) increased volume; (iii) faster retrieval of information.
(b) Control (i) greater accuracy and consistency; (ii) better security.
(c) Communication (i) enhanced communication; (ii) integration of business areas.
(d) Cost (i) monitor costs; (ii) reduce costs.
(e) Competitiveness (i) lock in customers; (ii) lock out competitors; (iii) improve arrangements with suppliers; (iv) new product development.

Sources of project requests (pp.41–42). "Mainly four primary sources": three inside (department managers, senior executives, systems analysts) and one outside (government agencies / outside groups).

Source Course description
Department managers (3.2.1) Deal with day-to-day activities; unhappy with the time staff take or with duplication of work; persuade higher authority to approve a computer-based system for office administration.
Senior executives (3.2.2) Presidents, vice-presidents; more information about the organisation and broader responsibilities; their requests "carry more weight and are generally broader in scope".
Systems analysts (3.2.3) Find areas for projects and write a proposal themselves or encourage a manager to do so (example: a slow library information system).
Outside groups (3.2.4) New external demands, e.g. government contractors must use a special cost-accounting system with government-stipulated features.

Managing project review and selection: three committees (pp.43–44). "Someone must decide which requests to pursue and which to reject."

Committee Members and role Notes
Steering committee (3.3.1) Key managers from various departments plus members of the information systems group; receives and evaluates requests; often needs a preliminary investigation for more information. One of the most common methods; projects are treated as business investments; decided on cost, benefit and feasibility within the limits of IS technology.
Information systems committee (3.3.2) Managers and analysts in the information systems department; approves or disapproves projects and sets priorities. Used when major equipment decisions or long-term development commitments are needed; senior executives share the final decision.
User-group committee (3.3.3) The users themselves: each department hires its own analysts and designers who select and develop projects. Takes burden off the development group but can leave users with defective or poorly designed systems; "success rate ... is not very encouraging".

The project request (3.3.4, p.44). A request form should contain: What is the problem? What are the details? How significant is it? What does the user feel is the solution? How will information systems help? Who else knows about this and could be contacted? (six questions). The committee reviews proposals and selects those most beneficial; a preliminary investigation is often requested.

Preliminary investigation (3.4, pp.44–45). "The first step in the system development life cycle." Its purpose is to evaluate project requests. "It is not a design study nor does it include the collection of details to describe the business system in all respect." Analysts should: clarify and understand the request; determine the size of the project; assess costs and benefits of alternative approaches; determine technical and operational feasibility of alternatives; report findings with recommendations on acceptance or rejection (five objectives).

Conducting the investigation (3.4.1, pp.45–46): three primary methods. (1) Reviewing organization documents (organisation charts, written operating procedures); (2) on-site observations (see the office, workload, methods and facilities, "get as close as possible to the real system"); (3) conducting interviews (documents and observation tell how the system should operate; interviews give user views and details showing whether assistance is merited economically, operationally and technically).

Testing project feasibility (3.4.2, pp.46–47). "Three important tests": operational, technical, economic. "To be judged feasible, a proposal ... must pass all these tests."

Feasibility Questions the course asks
Operational (p.46) Will the system work when installed? Is there sufficient support from management and users? Are current business methods acceptable to users? Have users been involved in planning and development (early involvement reduces resistance)? Will the system cause harm or poorer results, or lower staff performance?
Technical (p.47) Does the necessary technology exist (and can it be acquired)? Does the proposed equipment have the capacity to hold the data? Can the system be upgraded if developed? Are there technical guarantees of accuracy, reliability, ease of access and data security?
Economic (p.47) "Financial benefits must equal or exceed the costs." Estimate: the cost to conduct a full systems investigation; the cost of hardware and software for the class of application; the benefits (reduced costs or fewer costly errors); the cost if nothing changes.

Handling infeasible projects (3.4.3, pp.47–48). Requests that do not pass all tests "are not pursued further, unless they are modified and re-submitted as new proposals". The committee may combine the workable part with another feasible proposal. Sometimes the investigation shows that "improvements in management and supervision, not the development of information systems" are the real solution.

Problem classification and definition (3.5, pp.48–52). Questions: What is the problem? How complex? Likely causes? Why important to solve? Possible solutions? Expected benefits? (six). Seven common problem types (pp.48–51): the manager's budgeting complaint (p.48) shows reliability, accuracy, timeliness, validity and economy; capacity and throughput are also common.

Problem Definition and notes
(a) Reliability (p.49) Procedures work some but not all of the time, or the same procedure gives different results. Improved by software tests showing two runs give identical results, low-failure equipment, monitoring processing schedules. (Payroll, rail and airline reservation systems need it.)
(b) Validity (p.49) The system produces invalid results: it may be highly reliable and work all the time yet draw incorrect conclusions (e.g. report says demand is increasing when it is decreasing). Arises when the environment changes and the software is not updated.
(c) Accuracy (pp.49–50) Processing is error-prone (e.g. 1 % of posted transactions wrong). Inaccuracy can make the whole system look unreliable and invalid. Routine transaction-based manual procedures suit computerisation because the computer is far more accurate than people, provided the software is right.
(d) Economy (p.50) Existing methods of transmitting, processing and storing information are very costly (course example: paperwork per purchase order N25, cut to N8). Danger: the naive assumption that the computer will remove the cause; ask "Is the project worth doing?" and estimate the return on investment.
(e) Timeliness (p.50) Information is available but cannot be retrieved when and where needed; concerns transmission more than processing or storing.
(f) Capacity (p.51) A system component is not large enough (processing, transmission or storage), common at peak periods and in rapidly growing organisations. Example: two people doing the work of six. A quick-fix can create a worse problem later.
(g) Throughput (pp.51–52) The reverse of capacity: the efficiency of the system; capacity is high and production low. Example: six people doing the work of two.

Evaluating the problem (3.5.2, p.52). After identifying a problem: why is it important to solve, what are the possible solutions, what benefits can be expected. Sometimes the analyst recommends no project (example: a new supervisor who has learned from his mistakes); the analyst prepares a solutions table listing possible solutions and expected benefits. Rule of thumb: analysts spend 75 % of the project-definition phase of analysis defining the problem and 25 % evaluating and documenting findings. Feasibility entails two questions: "Can something be done?" and "Should it be done given a particular set of circumstances?" (mountain and forty-foot rope example, pp.52–53).

Sources of problems/opportunities (3.5.3, p.53): a new product, plant or branch; a new market or new process; failure of an existing system; inefficiency of an existing system; structural error in the existing system. Problem identification and definition (3.5.4, p.53): scan the performance of the system; the information supplied and its form; the economy of processing; the control of information processing; the efficiency of the existing system; the security of data and software; the security of equipment and personnel.

Code and worked examples

Example 1 — Name the problem (Course example p.48, the budgeting manager). The manager says: "Our current [system] seems to vary in quality from one month to the next. Besides, reports are often late, have errors, and contain misleading information. Why we must spend a fortune simply trying to keep the system up and going."

Phrase Problem type
varies in quality from one month to the next Reliability
reports are often late Timeliness
have errors Accuracy
contain misleading information Validity
spend a fortune keeping it going Economy

Example 2 — Capacity or throughput? (Course examples pp.48–49 and 51; the last row is mine).

  • Two people trying to do the work of six: capacity (the component is too small).
  • Six people doing the work of two: throughput (capacity is high, production low).
  • Equipment breaks down repeatedly and output is low: throughput (the course notes it can be easier to spot than to treat, p.52).
  • A warehouse that overflows during the December rush: capacity (peak-period storage).

Example 3 — The programming team (Course example p.51). Five programmers, 10,000 lines of code assigned, 30 days, 8 hours a day.

  1. Hours expended = 5 × 30 × 8 = 1200 hours.
  2. The course's rate: 6000 lines ÷ 1200 hours = 5 lines of code per hour per programmer.
  3. Share of the job done = 6000 ÷ 10 000 = 60 % (if the course's "600" were right, the rate would be 600 ÷ 1200 = 0.5 lines per hour and the job 6 % done; see Errata).
  4. Conclusion the course draws: capacity is high, output low, so a problem of throughput.

Example 4 — Economy of a purchase order (Course example p.50). Cost per order falls from N25 to N8. Saving per order = 25 − 8 = N17, i.e. 17 ÷ 25 = 68 %. (My illustration: at 1,000 orders a month that is N17,000 a month. The course gives no volume; only the N25 and N8 are the course's.) Before approving the project the analyst still asks "Is the project worth doing?" by checking return on the investment.

Example 5 — Which feasibility test? (Course questions pp.46–47; classification is mine).

Question Test
Is there sufficient support from management and users? Operational
Can the system be upgraded if developed? Technical
What is the cost if nothing changes? Economic
Will staff performance fall after implementation? Operational
Does the proposed equipment have capacity to hold the data? Technical
Do benefits (reduced costs, fewer errors) equal or exceed costs? Economic

Example 6 — A filled project request (structure from p.44; content is my illustration).

  1. What is the problem? Fee receipts take two days to reach the ledger.
  2. Details: handwritten receipt books, manual posting.
  3. How significant? About 300 receipts a week; frequent posting errors.
  4. What does the user feel is the solution? A computer-based receipting system.
  5. How will information systems help? Faster posting, fewer errors, quick retrieval.
  6. Who else knows about this? The head of accounts and the bursar.

Common mistakes

  • Four sources, three inside. Department managers, senior executives, systems analysts are inside; outside groups (government agencies) are the fourth.
  • Which committee does what. Steering = key managers plus IS group, projects as business investments. Information systems committee = IS managers and analysts, sets priorities. User-group committee = users select and develop themselves.
  • Reliability vs validity vs accuracy. Reliable but invalid is possible: the system works every time yet draws wrong conclusions. Inaccuracy = errors in processing. Reliability = consistent operation and results.
  • Capacity vs throughput. Capacity: the component is too small (two doing six people's work). Throughput: efficiency, plenty of capacity but low output (six doing two people's work).
  • Preliminary investigation is not a design study. It evaluates requests; it does not describe the business in full.
  • Unit 2's feasibility rule vs Unit 3's. Unit 2 says a proposal must pass all tests; Unit 3 (p.59) says a project "will not be rejected simply because of operational infeasibility" though it shapes the recommendation. For an MCQ quoting Unit 2, "must pass all" is the answer.

Errata

Page Course says Correct Note
p.51 10,000 lines assigned; "completed 600 usable lines of code" but then "5 lines of code per hour (6000 lines divided by 1200 hours)" 6000 usable lines gives 5 lines per hour; 600 lines would give 0.5 The two figures in the same example disagree; 6000 ÷ 1200 = 5 is the consistent one.

Practice questions

  1. (SAE i) Name the five main reasons for system projects: capability, control, communication, cost and ______.
  2. (SAE ii) Information systems add capability in three ways: improved processing speed, increased volume and ______.
  3. (SAE iii) How many primary sources of project requests does the course name? A. 2 B. 3 C. 4 D. 5
  4. (SAE iv) A department manager asks for a computer-based system because staff are slow or ______ work.
  5. (SAE v) Which of these is a communication reason for a systems project? A. Lock in customers B. Integration of business areas C. Better security D. Monitor costs
  6. The committee made up of key managers from various departments plus members of the information systems group is the ______ committee.
  7. How many questions does the course say a project request form should answer? A. 4 B. 5 C. 6 D. 7
  8. A system works every time but its reports wrongly say demand is rising. The problem is: A. Reliability B. Validity C. Capacity D. Timeliness
  9. Six people doing the work of two is a problem of ______.
  10. Analysts spend ______ % of the project-definition phase defining the problem and ______ % evaluating and documenting.
Answersshow
  1. Competitiveness (pp.40–41).
  2. Faster retrieval of information (p.40).
  3. C; department managers, senior executives, systems analysts, plus outside groups (p.41).
  4. Duplicating (duplication of) work (p.42).
  5. B; enhanced communication and integration of business areas (p.41). A is competitiveness, C control, D cost.
  6. Steering (p.43).
  7. C; six questions (p.44).
  8. B; the system is reliable but produces invalid results (p.49).
  9. Throughput (p.49).
  10. 75 and 25 (p.52).

Module 1 · Unit 3 of 4

Unit 3 — Feasibility Study

What you must know

  • Define feasibility and feasibility study, and say what the report is used for (p.57).
  • List the seven types of feasibility and define each (pp.57–60).
  • List the eight steps in feasibility analysis in order (p.60).
  • Explain the four types of benefits and the three pairs of cost/benefit categories (pp.62–64).
  • Give the five parts of the course's definition of cost-benefit analysis (p.65).
  • Describe fact-finding methods (interview, questionnaire, observation, documents) and the four questionnaire formats (pp.65–68).
  • Draw a context DFD and the analysis DFD steps for the accounts payable example, and describe the final feasibility report and meeting (pp.70–75).

Key ideas

Intuition. Feasibility asks two things: can it be done and should it be done. A feasibility study answers them before money is committed, and ends in a written and oral report.

Feasibility and feasibility study (p.57). "Feasibility is the determination of whether or not a project is worth doing. The process followed in making this determination is called a feasibility study." It determines "if a project can and should be taken". Once feasible, the analyst prepares the project specification that finalises requirements. Studies are done within tight time constraints and "normally culminate in a written and oral feasibility report", whose contents are the basis for deciding whether to proceed, postpone or cancel the project. A feasibility study is carried out to select the best system that meets performance requirements (pp.56–57).

Seven types of feasibility (pp.57–60): "distinct, but inter-related".

# Type Definition and points
1 Technical (p.58) Specifying equipment and software that will satisfy the user requirement. Needs may include: outputs in a given time; response time under certain conditions; volume of transactions at a particular speed; communicating data to distant locations. Configuration matters more than the make of hardware (how many workstations, how interconnected, I/O speeds, print quality); this becomes the basis of the tender document. Pursue two or three configurations at different levels of ambition and cost. "Technical feasibility generally is the most difficult to determine."
2 Operational (pp.58–59) Mainly human, organisational and political aspects: what changes will be brought; what organisational structures are disturbed; what new skills are required and can existing staff be trained. Projects are generally not rejected simply for operational infeasibility, but it affects the nature and scope of recommendations. Done by a small group familiar with IS techniques.
3 Economic (p.59) "The most frequently used technique for evaluating the effectiveness of a proposed system", more commonly cost/benefit analysis: determine expected benefits and savings and compare with costs. If benefits outweigh costs, decide to design and implement; otherwise further justification or alternatives are needed. An ongoing effort that improves in accuracy at each phase of the life cycle.
4 Social (p.59) Whether the project will be acceptable to the people; the probability of acceptance by the group directly affected.
5 Management (p.59) Whether the project will be acceptable to management; if management gives negligible support the analyst sees it as non-feasible.
6 Legal (p.59) Whether the project infringes known Acts, Statutes and pending legislation.
7 Time (p.60) Whether the project can be implemented fully within a stipulated time frame; if it takes too much time it is likely to be rejected.

Steps in feasibility analysis: eight (3.3.1, p.60).

  1. Form a project team and appoint a project leader.
  2. Prepare system flowcharts.
  3. Enumerate potential proposed systems.
  4. Define and identify characteristics of the proposed system.
  5. Determine and evaluate performance and cost effectiveness of each proposed system.
  6. Weigh system performance and cost data.
  7. Select the best proposed system.
  8. Prepare and report the final project directive to management.

Analysing systems data (3.3.2, p.60). "Systems analysis is fact finding followed by analysis of the facts." Data analysis is a prerequisite for cost/benefit analysis. The analyst develops a systems profile of each application area: operating characteristics such as frequency of occurrence, volume of work, error rate.

Identifying design requirements (3.3.3, pp.61–62). Requirements are determined by comparing current performance with the objectives for acceptable performance. The new system should have: (a) greater speed of processing; (b) effective procedure to eliminate errors; (c) better accuracy; (d) faster retrieval of information; (e) integration of data; (f) larger capacity of storing data with reduced cost. Alternatives are screened for economic, technical and operational feasibility. A computer's role: calculation, storage and retrieval, summarising, sorting, classification and communication of data (p.61).

Cost/benefit analysis (3.4, pp.62–65).

Four types of benefits (p.62–63):

Benefit Meaning Course example
Cost-savings Reductions in administrative and operational costs. Smaller clerical staff in future.
Cost-avoidance Eliminate future administrative and operational costs. No need to hire additional staff in future.
Improved-service-level Performance of a system improved by a new computer-based method. Registering a student in fifteen minutes instead of an hour (the course mislabels this; see Errata).
Improved-information Computer-based methods give better information for decisions. A system reporting the fifty most-improved customers by increase in sales.

Costs and benefits are classified three ways (p.63): tangible or intangible; direct or indirect; fixed or variable. "Costs ... are expenses, outlays or losses arising from developing and using a system. But the benefits are the advantages received."

Category Meaning (course) Course examples
Tangible (p.63) Can be measured; outlay of cash for an item or activity; known, estimated quite accurately. Cash outlays; completing jobs in fewer hours; error-free reports (benefits).
Intangible (p.63) Known to exist but financial value cannot be exactly measured; only an approximation. Employee morale affected by the new system; more satisfied customers; improved corporate image (benefits).
Direct (p.64) Directly associated with a system and attributable to a given project. Purchase of floppy disks for N400 (cost); a system processing 30 % more transactions per day (benefit).
Indirect (p.64) Not directly associated with a specific activity; "overhead expenses". Indirect benefits are a by-product of another system. Space, maintenance of the computer centre, heat, light, air-conditioning (costs); a sales-calls tracking system that also gives competition information (benefit).
Fixed (p.64) Constant regardless of use; sunk costs, "once encountered, they will not recur". Purchase of computer-centre equipment, insurance, software purchase (costs); staff reduced by 20 % (benefit).
Variable (p.64) Incurred regularly; generally proportional to work volume; continue while the system operates. Computer forms vary with the amount of processing (cost); time saved per borrower in a library system varies with the number of borrowers (benefit).

Definition of cost-benefit analysis (3.4.4, p.65): five parts. (i) the method by which we find and estimate the value of the gross benefits of a new system specification; (ii) find and determine the increased operating costs associated with those gross benefits; (iii) subtract these operating costs from the gross benefits to arrive at the net benefit; (iv) find and estimate the monetary value of the development costs that produce the benefits; (v) show the time-phased relationship between net benefits and development costs as they relate to cash flow, payback on investment and time-in-process, taking account of factors such as inflation: "the calculation of actual net benefit as cash flowback over time".

Fact finding (3.5, pp.65–70). "Learning as much as possible about the present system." The analyst: interviews personnel; prepares questionnaires; observes the current system; gathers forms and documents in use; determines the flow of data through the system; clearly defines the system requirements.

Method Course notes
Interviewing (pp.65–66) Study the organisation chart first; preliminary interviews with key personnel, then detailed interviews with those who operate the system (they may fear change). Questions: Who is involved with what you do? What do you do? Where? When? Why do you do it the way you do? How? Do you have suggestions for change? (seven). "It is the users' system, not the analyst's."
Questionnaires (pp.66–67) Economically gather data from large or small groups; results quickly tabulated; need planning and usually more than one draft. Questions should be short, easy to understand, unbiased, non-threatening, specific; test with one or two outsiders. Best for many people, the same questions for everyone, or outsiders such as suppliers.
Observation (pp.67–68) Follow a transaction (e.g. invoices) through the system; the analyst stays outside the flow; people observed "work more efficiently and at higher speeds to impress". Also visit a similar installation (etiquette: appointment, research, know what to see, thank-you letter); use vendors' information with care because vendors promote their products.

Four questionnaire formats (p.67): (i) multiple choice, ideal for computer tabulating; (ii) open ended, answers in the respondent's own words; (iii) rating, like multiple choice but respondents rate satisfaction; (iv) rank, prioritise from high to low or on a percentage basis. Analysts usually mix formats and add follow-up questions.

Fifteen questions to ask a software vendor (pp.68–70): range of products; decision support systems; in-house development; online; debugging and testing; updates; flexibility/adaptability; history/performance; other customers; security; networking; training support; in-house specialists; special features; upgrading.

Determination of DFD (3.5.4, pp.70–73). Armed with interviews, questionnaires and observation, the analyst describes the current system in narrative form, with a data-flow diagram or a system flowchart. A context DFD "defines the system under study in a general form" and "does not show any detail", placing a boundary around the system: "this is what will be examined, nothing more and nothing less". Good for management, whose interest is general.

  • Accounts payable (AP) context DFD (Fig 3.1, p.71): inputs: packing slips, invoices, checking account balances, payment notifications; outputs: reports to management, cheque to suppliers (p.70).
  • The detailed activities add up to six major activities (Fig 3.2, p.72): generation of reports, ordering of stock, printing of cheque, posting of accounts, reconciliation of bank statements, authorization of payment.
  • To draw the analysis DFD (seven steps, p.72): 1. look at the system from the inside to the outside; 2. identify the activities; 3. locate the data flows; 4. show relationships between activities; 5. find internal inputs or outputs within the system; 6. level complex processes into simpler ones; 7. look for duplication of data flows or data stores (files).
  • Collect sample documents: invoice data (supplier name, address and phone; invoice number; date; terms; amount), packing slip (supplier name, shipping date, date goods received, freight charges, invoice number; a carbon copy of the invoice without money value), ledger (supplier data and purchase/payment history, eight items) and cheque (invoice number, cheque number, amount, payment date) (pp.73–74).

New system and the final report (3.5.5, pp.74–75). First list alternative solutions, "from doing nothing to installing a fully computerized AP system". The AP example has four alternatives: (i) do nothing; (ii) hire more staff, partially automate, remain essentially manual; (iii) purchase AP software from an outside supplier; (iv) design, program and install a customised AP system. The final report, called the feasibility study, starts with a restatement of the problem and its importance, then the study's objectives, a review of findings, tallies of expected costs and savings, and the recommendations. Management then meets to choose a course of action. Rules for leading the meeting (p.75): (i) never read the study aloud, summarise it; (ii) use visual aids; (iii) if appropriate, demonstrate equipment or software. After the meeting management notifies staff by memo even if nothing changes. After a decision to proceed, analysis ends and design begins; all memoranda, questionnaires, interview documents, forms, DFDs and reports are organised into one file, the analysis documentation.

Code and worked examples

Example 1 — Which of the seven feasibilities? (course definitions pp.58–60; scenarios are mine).

Scenario Feasibility
Can the planned servers handle 500 transactions a minute? Technical
Will the staff accept the new procedures, and can they be trained? Operational
Do the expected savings exceed the cost? Economic
Will the public accept the new online service? Social
Will the managing director support it? Management
Does storing the data this way breach a data-protection law? Legal
Must the system be live before the new term starts in nine months? Time

Example 2 — Classify the costs and benefits (Course examples pp.63–64).

Item Tangible/intangible Direct/indirect Fixed/variable
Floppy disks for N400 Tangible Direct not stated
Heat, light and air-conditioning of the computer centre Tangible Indirect not stated (the course: hard to apportion to one report)
Purchase of equipment for the computer centre Tangible Direct Fixed (constant whether or not used extensively)
Computer forms Tangible Direct Variable (proportional to processing or report length)
Staff reduced by 20 % Tangible Direct Fixed benefit
Two minutes saved per library enquiry Tangible Direct Variable benefit (varies with number of borrowers)
Competition information from a sales-call tracking system Intangible Indirect (by-product) —
Employee morale hurt by the change Intangible cost — —

(Each example appears in the course under one category; "not stated" and the other columns are my reading, not marked by the course.)

Example 3 — Net benefit and payback (the five-part definition p.65; the numbers are my illustration, and the payback formula is the standard one, not stated in the course). Suppose gross annual benefits are N480,000 and added annual operating costs are N230,000, and development cost is N600,000.

  1. Net annual benefit = gross benefits − operating costs = 480,000 − 230,000 = N250,000 (part iii).
  2. Development cost = N600,000 (part iv).
  3. Payback period = development cost ÷ net annual benefit = 600,000 ÷ 250,000 = 2.4 years (the "time-phased relationship", part v). If the study horizon were two years the project would not have paid back; the analyst would look at further justification or alternatives (p.59).

Example 4 — Which benefit type? (Course examples pp.62–63).

  • "Fewer clerks needed next year" → cost-savings.
  • "No extra clerks will have to be hired when volume grows" → cost-avoidance.
  • "Student registration in 15 minutes rather than an hour" → improved-service-level (the course text says cost-avoidance; see Errata).
  • "Report on the fifty customers whose sales improved most" → improved-information.

Example 5 — The AP context DFD and analysis DFD as text (Course Figures 3.1 and 3.2, pp.71–72; flows read from the figures, redrawn as tables). Context DFD (one process, the AP system, inside a boundary):

External entity Flows into AP Flows out of AP
Supplier Packing slips, invoices Check, purchase order
Bank Checking account balances —
Management Payment authorization Reports

Analysis DFD (Fig 3.2 as drawn shows three process bubbles): Order stock (sends purchase order requisition to suppliers; takes management reports), Post accounts (takes packing slips and invoices), Print checks (takes payment authorisation from management and checking account balances from the bank; sends the check to the supplier). The text lists six major activities (generation of reports, ordering of stock, printing of cheque, posting of accounts, reconciliation of bank statements, authorization of payment); the figure shows only three as bubbles (see Errata).

Example 6 — Choose among the four AP alternatives (course list p.74; the screening is mine and uses only the course's three tests).

Alternative Technical Economic Operational
(i) Do nothing Trivially feasible Costs of not changing continue (p.30: the cost of not creating the system counts) No change, no resistance, but the problem remains
(ii) More staff, partial automation Feasible Recurring staff cost grows Familiar to staff
(iii) Buy AP software Check against the 15 vendor questions (pp.68–70) Purchase plus training Needs staff training
(iv) Custom AP system Most demanding Highest development cost Users must be involved from the start

The analyst weighs performance and cost (step vi), selects the best (step vii) and reports to management (step viii).

Common mistakes

  • Seven types, not three. Unit 3 lists technical, operational, economic, social, management, legal, time. Units 1 and 2 list only the first three (in different orders).
  • Technical = configuration. The course says configuration, not the make of hardware, is given more importance (p.58).
  • Fixed does not mean "never changes in amount". "Fixed costs are sunk costs; once encountered, they will not recur" (p.64). Variable costs recur and track work volume.
  • Tangible vs direct. Tangible/intangible is about measurability; direct/indirect is about attribution to one activity. Indirect costs such as heat and light are still tangible.
  • Cost-savings vs cost-avoidance. Savings reduce existing costs (smaller clerical staff in future); avoidance eliminates future costs (no additional hires).
  • Questionnaire formats. Multiple choice, open ended, rating, rank. Rating is multiple-choice-like satisfaction scoring; rank asks respondents to order or give percentages.
  • Context DFD vs analysis DFD. The context DFD shows no internal detail and fixes the boundary; the analysis DFD is levelled into activities.

Errata

Page Course says Correct Note
p.57 "In subsection 3.3.1 of unit 1" Unit 1 section 3.2.1 (preliminary investigation, p.29) Unit 1 has no 3.3.1 on feasibility; 3.3.1 is a software-crisis heading.
p.62 Registering a student in fifteen minutes rather than an hour is "an example of a cost-avoidance benefit" Improved-service-level benefit The sentence is the example for the improved-service-level paragraph, yet it names cost-avoidance.
p.63 "If the project is evaluated on a purely intangible basis, benefits exceed costs ... Hence ... evaluated on the basis of intangible benefits alone" Tangible basis / tangible benefits alone The argument is that intangibles cannot be measured; the text twice says "intangible" where "tangible" is meant.
p.72 Six major activities listed for Fig 3.2 Fig 3.2 shows three process bubbles (Order Stock, Post Accounts, Print Checks) Reports, bank reconciliation and authorisation appear only as flows or entities.
pp.70–71 Text: inputs to AP include "payment notifications"; outputs are reports and cheque Fig 3.1 shows payment authorization (from management) as an input and a purchase order to suppliers as an output The figure has more flows than the text lists.
p.73 Invoice data item "4. Invoice" Item is incomplete (probably "Invoice item" or amount line); item 6 is "Amount of invoice" List as printed has only the word "Invoice".

Practice questions

  1. (SAE i) "The determination of whether or not a project is worth doing" is called ______.
  2. (SAE ii) How many types of feasibility does Unit 3 list? A. 3 B. 5 C. 7 D. 8
  3. (SAE iii) Technical feasibility is concerned with specifying ______ and ______ that satisfy the user requirement.
  4. (SAE iv) Which is a point considered in operational feasibility? A. Tender document B. What new skills will be required C. Net benefit D. Statutes
  5. Which feasibility asks whether a project infringes known Acts and Statutes? A. Social B. Management C. Legal D. Time
  6. (TMA 1) The four types of benefits are cost-savings, cost-avoidance, improved-service-level and ______.
  7. Costs that cannot be exactly measured in money are called ______ costs.
  8. The purchase of equipment for a computer centre is a ______ cost. A. variable B. fixed C. indirect D. intangible
  9. How many steps are in the feasibility analysis (3.3.1)? A. 6 B. 7 C. 8 D. 15
  10. A context DFD ______. A. levels every process B. shows no internal detail and fixes the boundary C. lists the 15 vendor questions D. replaces the feasibility report
Answersshow
  1. Feasibility (the process is the feasibility study) (p.57).
  2. C; seven (p.57).
  3. Equipment and software (p.58).
  4. B; also what changes, what structures are disturbed, can staff be trained (p.58).
  5. C; legal feasibility (p.59).
  6. Improved-information (p.62).
  7. Intangible (p.63).
  8. B; fixed (p.64).
  9. C; eight steps (p.60).
  10. B; (p.71).

Module 1 · Unit 4 of 4

Unit 4 — System Requirement Specifications and Analysis

What you must know

  • Define a DFD, name its four symbols, and state its seven construction rules and the balancing rule (pp.78–80).
  • Compare the DeMarco, Gane-Sarson and Weinberg notations for DFD forms (pp.80–82).
  • Define a data dictionary, its four symbols, four rules, entry format and the types (integrated/stand-alone; passive/active/in-line) (pp.82–90).
  • Explain HIPO, its two diagrams VTOC and IPO, how a VTOC is numbered and decomposed (pp.90–96).
  • Build a decision table (stubs, entries, 2^N rules, redundancy dash) and name limited, extended, mixed and open-ended tables (pp.96–101).
  • Describe dialogue trees and decision trees (pp.101–104).
  • Describe Warnier-Orr diagrams and Nassi-Shneiderman charts and their advantages and disadvantages (pp.105–110).

Key ideas

Intuition. Unit 4 is the toolbox for writing down a system's requirements. The tools fall into groups: data: DFD, data dictionary, Warnier-Orr; structure: HIPO (VTOC, IPO); logic: decision tables and trees, N-S charts.

Data flow diagram (DFD) (3.1.1, p.78). "Graphical description of a system's data and how the processes transform the data." Unlike detailed flowcharts, DFDs "do not supply detailed descriptions of modules but graphically describe a system's data and how the data interact with the system."

Symbol Meaning (p.78)
Arrow A data flow: data in motion, "a pipeline through which information flows".
Circle A process: converts data into information (like the rectangle in flowcharts).
Open-ended box A data store: data at rest, "a temporary repository of data".
Square A source (originator) or destination of system data.

Seven rules governing DFD construction (pp.78–79).

  1. Arrows should not cross each other.
  2. Squares, circles and files must bear names.
  3. Decomposed data flows must be balanced (all flows on the decomposed diagram reflect flows in the original diagram).
  4. No two data flows, squares or circles can have the same name.
  5. Draw all data flows around the outside of the diagram.
  6. Choose meaningful names for data flows, processes and data stores: "strong verbs followed by nouns".
  7. Control information such as record counts, passwords and validation requirements is not pertinent to a DFD.

Decomposition and balancing (pp.79–80). If too many events occur at one point, a process (circle) is decomposed into child circles; parent and child form a parent-child relationship (the child circle in Fig 3.2 belongs to the parent in Fig 3.1). The balancing feature "builds in an error-detection system other tools lack": if the parent shows three inputs and two outputs, the levelled child diagrams together must show three inputs and two outputs; an imbalance means an error in the parent or the child.

Charting tools for DFDs (3.1.2, pp.80–82). "The data flow diagram is the core specification in this method." Figure 3.3 shows the charting forms: (a) square box = external source or receiver of data; (b) transform bubble; (c) line arrow; (d) data store; (e) Boolean logic.

Topic Course statement
Transform bubble (p.81) Two variations: the circle (used by Victor Weinberg and Tom DeMarco, the better known) and the rectangular bubble (Chris Gane and Trish Sarson) because of the perceived need for more information than the bubble holds. DeMarco, "the purist", holds that bubble content is just one verb and a noun, because the aim is to partition into levelled transforms.
Line arrow (p.81) Carries the data flow; all lines must be identified by their data. Three points of view: (1) Boolean logic (AND, INCLUSIVE OR, EXCLUSIVE OR) for multiple lines in or out is offered by Weinberg; DeMarco advises against it. (2) DeMarco draws a curved line (a sense of data in motion); Weinberg and Gane-Sarson use straight lines and right angles. (3) Break wordy flows into several briefly named flows.
Data store (p.82) Open rectangle or two parallel lines (database, file, Kardex, phone book); Gane and Sarson show two parallel lines joined at one side.

"A system of any complexity whatever is shown with the bubble, line, data store rectangle and external box" (p.82).

Data dictionary (3.2, pp.82–90). "A data dictionary defines each term (called a data element) encountered during the analysis and design of a new system." Data elements can describe files, data flows or processes. Example (p.82): Vendor name and address = Vendor name + Street + City + State + Pin + Phone + Fax + e-mail.

Major symbols (p.83): = equivalent to; + and; [ ] either/or; ( ) optional entry.

Four rules for data dictionary entries (p.83).

  1. Words are defined to stand for what they mean, not the program's variable names (use CLIENT_NAME, not ABCPQ or CODE06); capitalisation helps.
  2. Each word must be unique.
  3. Aliases (synonyms) are allowed when two entries have the same meaning (vendor number = customer number), only when absolutely necessary.
  4. Self-defining words should not be decomposed. (The course also shows decomposing a definition: Vendor name = [Company name | Individual's name]; Company name = (Contact) + Business name; Individual's name = Last name + First name + (Middle initial).)

Entry format (pp.83–84), e.g. VENDOR_NUMBER: data element name; aliases (none); description (unique identifier for vendors in the accounts payable system); format (alphanumeric, six characters); data flows where found (vendor master, AP open item ...); reports that use it. Dictionaries seldom hold passwords or statistical facts; they define words and terms. Data Dictionary systems (DDS) are packages that maintain dictionaries by computer, tracking each term, its definition, which systems use it, aliases, number of uses and size (p.84).

Data dictionary types (3.2.4, pp.85–86). Two kinds by tie to a DBMS, three by function (Fig 3.5).

Classification Type Course description
By relation to a DBMS (p.85) Integrated Related to one DBMS; global/organisation-wide only to the extent that data are under that DBMS; since few enterprises keep all data in one DBMS, the documentation (metadata) is local and fragmented.
Stand-alone Not tied to any DBMS (e.g. IBM DB-DC Data Dictionary, which has special features for IBM IMS but is still stand-alone).
By function (p.86) Passive Documentation only; can be manual; the documentation function is the most important, so passive is "not to be thought of negatively".
Active Supports documentation and program/operations development by exporting database and program data-storage definitions for languages such as COBOL and JCL. The IBM DB/DC dictionary is a stand-alone, active one.
In-line Active during program performing, e.g. transaction validation and editing; associated with DBMS products such as IDMS-R and TOTAL.

Make-up of data dictionaries (3.2.5, pp.86–90). A database consists of files; files consist of data groups (segments or records); data groups consist of data items (fields). Attribute information: primary or secondary key or attribute field; aliases; field type and size; status (test, system test or production); number of occurrences. The IBM-style item name has four elements, e.g. T,C,BALANCE-ON-HAND,0: T = status code (test-data database), C = subject code (COBOL, the primary programming language), the name, and 0 = occurrence number (p.87). Another example: P,,Provide-Materials,0: P = Production, two adjacent commas mean the subject code is not used, 0 = only this occurrence (p.89). The full-service (extended) dictionary is made of categories, subjects, relationships, attributes and descriptions (Fig 3.7, p.88); relationships run forward and reverse (Projects ACCOMPLISH Information Systems / Accomplished By). Its extensibility, the ability to define user categories, is "the heart of the full-service dictionary". The IBM data dictionary is six linked databases with eleven standard categories: DATA-BASE, SEGMENT, ELEMENT, PROGRAM COMMUNICATION BLOCK, IMS SYSTEM DEFINITION, APPLICATION SYSTEM, JOB, PROGRAM, MODULE, TRANSACTION, PSB (p.89).

HIPO (3.3, pp.90–96). Hierarchy plus Input Process Output: two diagrams, the Visual Table of Contents (VTOC) and the Input Process Output (IPO) chart; together they assist in designing programs and their functions.

VTOC IPO
Shows Overview: the system broken top-down into increasingly detailed levels; system name at the top, major functions on level 2, sub-functions on level 3 and below Inputs, processing and outputs of one module: the program logic
Reading Top down and left to right; modules in order of priority Lists activities, "not necessarily ordering them in the sequence they should occur"
Module naming Rectangle, two to four words, verb followed by object ("compute net pay") Header: system, module number, name, description, date, author; three columns: input (files used), process, output
Replaces An alternative to the system flowchart Replaces the program (detail) flowchart

Constructing a VTOC (3.3.1, pp.92–95). Assign all outputs (reports) to modules; name each by function (verb-object); number: system 1.0, first-level modules 1.1, 1.2 ...; decompose a module that does more than one task ("a module must be single purpose"): the AP check module has two tasks, the remittance advice (upper stub: date, number, discount, balance, total of each invoice) and the cheque (lower stub: cheque number, payment amount, vendor name and number), giving 1.1.1 and 1.1.2. "Decomposition ends when all modules are single purpose, single entry, and single exit." The analyst can stop at the third level; the programmer decomposes further. Programmers call VTOCs structure charts, hierarchy charts or tree charts, and code from the top left, finishing each level before moving down and right (p.95).

HIPO advantages (p.96): drawn or modified rapidly; convey the system graphically to non-computer people; standard symbols let a future analyst grasp it quickly; easy to estimate programming time per module, which simplifies assignments.

Decision tables (3.4.1, pp.96–101). Decision tables and trees pre-date computers; they "isolate many conditions and possible actions" and help ensure nothing is overlooked.

Part Position Meaning
Condition stub upper left List of all the necessary tests (yes/no questions, numbered), like the diamond in a flowchart or the IF in pseudo-code
Action stub lower left List of all the processes (actions) in the module, numbered
Condition entry upper right All yes/no permutations; each vertical column is a rule, numbered 1, 2, 3 ...
Action entry lower right X = action occurs, dot (.) = no action
Header top Title, date, author, system, comments

The chart is described as four sections plus the top header (Fig 3.17 is captioned "Five sections"). Number of rules = 2^N, N = number of conditions (p.96): 4 conditions → 16 rules, 6 → 64, 7 → 128, 8 → 256. Building the condition entry (pp.97–98): top row, half y then half n; each next row alternates in half the block size, so the pattern covers every combination without duplication; the last row alternates y n y n. Redundancy: where two or more rules give the same actions they can be combined by putting a dash (-) in the condition entry, meaning "irrelevant, can be ignored" (p.98).

Type of decision table Definition (pp.98–100)
Limited entry Lists a y or n response for each condition (Fig 3.19).
Extended entry Displays the values to be tested in the condition entry (e.g. End/More) (Fig 3.20).
Mixed entry Mixes values and yes/no in the condition entries (Fig 3.21).
Open-ended Permits access to another decision table (an action entry specifying an additional table) (Fig 3.22); used to make a table more readable or to split a large table (seven conditions = 128 rules).

Dialogue trees and decision trees (pp.101–104). Three tools for interactive dialogue design: dialogue trees, decision trees, picture-frame analysis (picture-frame analysis is a walkthrough of how screens will appear). A dialogue tree "maps the static and dynamic messages that take place between the computer and the user" (Fig 3.23, delete-a-record menu). An expanded tree (Fig 3.24) uses a non-restricted node (X): not restricted to a prescribed number of choices. A decision tree drops the dialogue detail and shows "the paths that are possible in a design following an action or decision by the user" (Fig 3.25: selecting 1, then M, then A shows the algebra menu); it gives "an overview of the flow of control". Decision trees turn a decision table into a diagram: read left to right, decisions result in a fork, every branch ends with an outcome; non-technical users find them easier than tables (Fig 3.26, p.104).

Warnier-Orr diagrams (3.5, pp.105–108). Named for Jean-Dominique Warnier and Kenneth Orr. Unlike VTOCs, pseudo-code or flowcharts (top down, then left to right) it reads left to right, then top to down. Uses brackets, circles, parentheses, dots and bars. Can show data-dictionary-type definitions (Fig 3.27) or program logic (Fig 3.28).

  • Sequence: a bracket groups related elements (Systems process = Analysis + Design + Development).
  • Repetition (iteration): a parenthesis left of the repeated series; a number inside gives how many times (0, End = repeat until the end).
  • Selection (alternation): a plus sign in a circle; a bar separates true (above) from false (below) (Fig 3.29).
  • Begin / Process / End parts are shown explicitly (Fig 3.29, AP stub-over-cheque).
  • Advantages (p.106 and p.108): single entry and single exit; supports the three control structures; few symbols; easy to construct, read and interpret; decomposes complex terms; describes the three logical constructs; analyses both actions and things; simplifies the definition and order of terms entered into the data dictionary.
  • Disadvantages: left-to-right, top-down construction (the opposite of other tools); focus on processing versus data flow; does not show relationships within and between data stores (p.108).
  • Decomposition ends when attributes can be defined by field length or physical storage, e.g. City_address = 2 (character) 12 (two to twelve characters), Telephone_extension = 4 digits (p.108); Employee_master_file = 1[employee_records]n (p.107).

Nassi-Shneiderman (N-S) charts (3.6, pp.109–110). An alternative to pseudo-code or program flowcharts; "much more compact than program flowcharts", with pseudo-code-like statements and the sequence, decision and repetition constructs; also called Chapin charts; developed by I. Nassi and B. Shneiderman in the early 1970s. Rectangles divided into halves with an angular line for selection, a horizontal rectangle for sequence, the word DOWHILE for iteration. Single entry, single exit, accommodate modules. Disadvantages: not as useful for conveying system flow as for detailing logic; difficult to maintain.

Tool Reads Best for
DFD — Data flow, sources and destinations (p.78)
Data dictionary — Defining every data element (p.82)
VTOC Top down, left to right Hierarchy of modules (p.90)
IPO — Inputs, processing, outputs of a module (p.92)
Decision table — Conditions and actions, nothing overlooked (p.96)
Decision tree Left to right Table as a diagram for non-technical users (p.104)
Warnier-Orr Left to right, then top to down Data-store contents and logic (p.105)
N-S chart — Compact program logic (p.109)

Code and worked examples

Example 1 — Balancing a parent and child DFD (Course Figures 3.1 and 3.2, pp.79–80, checked against the figures; text rendering is mine). Parent (Fig 3.1): process 3.0 Print AP Checks with

  • inputs: Sorted Invoice File (from process 2.0), Old Vendor Master (a data store) — 2 inputs;
  • outputs: New Vendor Master, Stub-over Checks to the entity Vendors, Disks with Files to the entity LIBRARY — 3 outputs.

Child (Fig 3.2): 3.1 Print Remittance Advice → 3.2 AP Checks, with

  • inputs entering the child: Sorted Invoices, Old Vendor Master — 2;
  • outputs leaving it: New Vendor Master, Stub-over Checks to Vendors, Disks with Files to LIBRARY — 3;
  • the arrow between 3.1 and 3.2 is internal and does not appear on the parent.

2 in + 3 out on both: the diagrams are balanced. Had the child shown a fourth output not on the parent, the course's rule says there is an error in one of the two (p.80).

Example 2 — Reading data dictionary notation (Course examples pp.82–83).

  1. Vendor name and address = Vendor name + Street + City + State + Pin + Phone + Fax + e-mail: "+" means all eight parts are present.
  2. Vendor name = [Company name | Individual's name]: either/or, so a vendor name is one or the other.
  3. Company name = (Contact) + Business name: Business name always; Contact optional.
  4. Individual's name = Last name + First name + (Middle initial): two valid forms, with or without the middle initial.
  5. Entry for VENDOR_NUMBER (p.84): aliases none; description "unique identifier for vendors in the accounts payable system"; format alphanumeric, six characters; listed under the data flows Vendor master, Accounts payable open item and so on.
  6. Rule check: the program might call the field CODE06; rule 1 says the dictionary term must be a meaningful name such as VENDOR_NUMBER.

Example 3 — Numbering a VTOC and reading an IPO (Course pp.92–95). VTOC for the AP cheque system:

Text
1.0 AP system
 ├── 1.1 Print AP cheque
 │     ├── 1.1.1 Print remittance advice   (upper stub)
 │     └── 1.1.2 Print cheque              (lower stub)
 └── 1.2 Print numeric vendor list  ...

The course's programmer example (p.95) codes in this order: 1.1, 1.1.1, 1.1.2, then 1.2, 1.2.1, 1.2.2, 1.2.3 (the course prints the second number as "1.2.1" a second time; see Errata). Rule: finish a level top-left before moving down and to the right. IPO for Finish Assembly (Course Fig 3.11, p.92), in the course's three columns:

Input Process Output
Frame; front-wheel assembly; rear-wheel assembly; seat; chain; handlebars 1. Bolt front- and rear-wheel assembly to frame; 2. Attach handlebars; 3. Attach chain; 4. Attach seat Completed bicycle

Example 4 — Counting and building decision-table rules (Course formula p.96, pattern pp.97–98; verified by script). Rules = 2^N: N = 2 → 4; N = 3 → 8; N = 4 → 16; N = 6 → 64; N = 7 → 128; N = 8 → 256. (The course's "2N" would give 8 for N = 4, which contradicts its own 16.) For 3 conditions use the course's method: top row half yes then half no; each following row halves the block size.

Text
Rule        1 2 3 4 5 6 7 8
Condition 1 y y y y n n n n
Condition 2 y y n n y y n n
Condition 3 y n y n y n y n

Course bicycle table (Fig 3.18, p.97), 2 conditions, 4 rules, 5 actions (4 × 5 = 20 action entries):

1 2 3 4
1. Last carton? y y n n
2. Hand brakes? y n y n
Open container . . X X
Stack parts . . X X
Assemble wheels . . X X
Finish assembly . . X X
End of assembly operations X X . .

Rules 1 and 2 give the same actions whatever the answer to condition 2, so combine them with a dash: y, - → End of assembly operations; likewise rules 3 and 4: n, - → the four assembly actions (p.98).

Example 5 — Tracing the AP cheque decision table and tree (Course Figures 3.19 and 3.26, pp.99 and 104; trace data are mine). Table conditions: C1 End of vendor master file? (the course prints "Hand of sorted invoice file?" for C2, a misprint for "End"; see Errata), C2 End of sorted invoice file?, C3 Do vendor numbers match? Active rules: rules 1 and 2 (C1 = y, C2 = y, C3 any): End of module. Rule 7 (n, n, y): read an invoice record, add amount to total, print invoice detail line. Rule 8 (n, n, n): read a vendor master record, print date line, print amount in words, print vendor name/address. Rules 3–6 have no X in the course's table.

Vendor master: V1, V2. Sorted invoices: (V1, 100), (V1, 50), (V2, 70). Start with V1 and the first invoice already read (the table has no "start" rule).

Step Current vendor / invoice C1 C2 C3 Rule Action
1 V1 / (V1,100) n n y 7 total = 100; print detail; read next invoice
2 V1 / (V1,50) n n y 7 total = 150; print detail; read next invoice
3 V1 / (V2,70) n n n 8 print date line, amount in words (150), vendor name/address; read vendor V2
4 V2 / (V2,70) n n y 7 total = 70; print detail; read next invoice (end of file)
5 V2 / end of invoices n y — 5 or 6 No action in the table as drawn

Reading note (mine, not examined): the table as printed has no action for rules 3–6, so the final vendor's date line and amount in words would not print; the course's N-S chart (Fig 3.33) prints them after the inner loop. The decision tree (Fig 3.26) says the same thing as rules 7, 8 and 1/2: Are vendor numbers same? → add amount to total, print detail line, read invoice record. Different → end of file → End module; record exists → print date line, print amount in words, print vendor name/address, read vendor record.

Example 6 — Dialogue tree, Warnier-Orr and N-S (Course pp.102, 107–108, 109; the control-sum trace is mine). Dialogue tree for the delete-a-record menu (Fig 3.23, p.102). Response 2 → DELETE A RECORD from the employee master file. Prompt EMPLOYEE NAME OK? If yes → SURE YOU WANT TO DELETE?; if no → NAME OK NOW?. Record is removed only if the name is OK (or OK now) and the user answers yes to SURE YOU WANT TO DELETE. Otherwise control shows "return to start", a loop back to the start of the tree. Warnier-Orr decomposition (pp.107–108). Employee_master_file = 1[employee_records]n; Employee_record = Employee number + employee_name_and_address + personnel_information + wage_and_salary_information + leave_and_pension_history + current_payroll_history (six parts); Employee_name_and_address = Employee name + employee_home_address + employee_plant_address + telephone_extension_number (four parts); Employee_home_address = ({PO_box_/apartment_number}) + street_address + city_address + state_address + Pin_code, where the PO box or apartment number is optional. N-S payroll chart (Fig 3.31, p.109) traced for 3 payroll records and cheque total 3. Enter cheque-number; set control sum = 0; enter cheque total (3); get the current payroll record. Until EOF = yes: increment controlsum (1, 2, 3); write paycheque; increment cheque-number; get next record. After EOF: Controlsum = Cheque total? 3 = 3, yes → print "Cheque total is ok"; if not equal → "Warning: Cheque total in error".

Common mistakes

  • Symbols. Circle = process, arrow = data flow, open-ended box = data store, square = source/destination. MCQs swap them.
  • Rule 7 of DFD. Control information (record counts, passwords, validation) is not shown on a DFD. Rule 3: decomposed flows must be balanced.
  • Data dictionary symbols. = equivalent to, + and, [ ] either/or, ( ) optional. Do not confuse ( ) optional with [ ] either/or.
  • Integrated vs stand-alone vs passive/active/in-line. The first pair is about tie to a DBMS; the second is about function. Passive = documentation only; active = exports definitions (COBOL, JCL); in-line = active during program running, validation and editing.
  • VTOC vs IPO. VTOC = hierarchy overview (replaces the system flowchart); IPO = logic of one module (replaces the program flowchart). HIPO = Hierarchy plus Input Process Output.
  • Decision table. Rules = 2^N. Stubs are on the left, entries on the right; X = action, dot = no action; dash = condition ignored. Limited = y/n, extended = values, mixed = both, open-ended = leads to another table. (Open-ended also means a questionnaire item answered in one's own words, p.100.)
  • Warnier-Orr reads left to right, then top to down (the opposite of VTOC); N-S charts are also called Chapin charts; DOWHILE marks iteration.
  • Warnier-Orr disadvantage. It does not show relationships within and between data stores; its focus is on processing, not data flow.

Errata

Page Course says Correct Note
pp.79–80, 82, 84–87, 94 Text refers to "Figure 4.1, 4.2, 4.3, 4.4, 4.5, 4.6, 4.7" and "figure 4.15" The figures are labelled 3.1–3.7 and 3.15 The unit was renumbered from 4.x to 3.x in the figure captions only; Unit 3 also has its own Figures 3.1 and 3.2.
p.83 "Vendor name = Company name, Individual's name" Vendor name = [Company name | Individual's name] (either/or) The course prints a comma where its own symbol list uses [ ] for either/or.
p.95 "print amount in words (1.2.1)" after "print date line (1.2.1)" 1.2.2 The three sub-modules are 1.2.1, 1.2.2, 1.2.3.
p.96 "Number of rules = 2 ^ N = 2N" 2^N 2N is wrong: the same paragraph says 4 conditions give 16 rules.
p.98 "all combinations with duplication" without duplication The pattern in the text has none.
p.99 Fig 3.19 condition 2 "Hand of sorted invoice file?" End of sorted invoice file? Matches condition 1 and the End/More entries of Fig 3.20.
p.99 Fig 3.20 and Fig 3.21: third row ("Do vendor numbers match?") entries are End/More A match test needs yes/no (or same/different) entries The row repeats the End/More values of the other rows.
p.105 "repeat the four bracketed elements until there are more bicycles to assemble" until there are no more bicycles Fig 3.28 iteration is (0, End).
p.105 "Technically the symbol for a bracket is '(' and a brace is '}'" A bracket is [ and a brace is { The printed sentence is garbled; the course then calls "{" a bracket.
p.110 Fig 3.33 (AP cheque N-S chart) outer loop "DOWHILE there are containers" There are invoice records (to process) Copied from the bicycle chart, Fig 3.32.

Practice questions

  1. In a DFD the circle represents a ______.
  2. Which DFD symbol stands for a source or destination of data? A. Arrow B. Circle C. Square D. Open-ended box
  3. How many rules does the course give for constructing a DFD? A. 4 B. 5 C. 7 D. 10
  4. A parent DFD shows three inputs and two outputs. The child diagrams together must show ______ inputs and ______ outputs.
  5. In a data dictionary, the symbol [ ] means ______. A. and B. equivalent to C. optional D. either/or
  6. (TMA 3 b) Which is one of the four rules for data dictionary entries? A. Each word must be unique B. Every word needs an alias C. Self-defining words must always be decomposed D. Use program variable names such as ABCPQ
  7. A data dictionary that only performs documentation is ______. A. active B. passive C. in-line D. integrated
  8. (TMA 4) HIPO consists of the ______ (VTOC) and the ______ (IPO) chart.
  9. A decision table has 4 conditions. It has ______ rules. A. 4 B. 8 C. 16 D. 32
  10. Which tool is "also called Chapin charts" and uses the word DOWHILE for iteration? A. Warnier-Orr diagram B. Nassi-Shneiderman chart C. VTOC D. Decision tree
Answersshow
  1. Process (a data conversion) (p.78).
  2. C; square (p.78).
  3. C; seven (pp.78–79).
  4. Three and two (p.80).
  5. D; either/or (p.83). = equivalent to, + and, ( ) optional.
  6. A; the four rules: meaningful names, unique words, aliases only when necessary, self-defining words not decomposed (p.83).
  7. B; passive (p.86).
  8. Visual Table of Contents and Input Process Output (p.90).
  9. C; 2^4 = 16 (p.96).
  10. B; Nassi-Shneiderman (p.109).

External resources

Each page below returned HTTP 200 when opened (checked 2026-11-22). They are general references; for the exam, the course's own wording above is what counts.

  • Data-flow diagram (symbols, levels, balancing): https://en.wikipedia.org/wiki/Data-flow_diagram
  • Systems development life cycle (phases and models): https://en.wikipedia.org/wiki/Systems_development_life_cycle
  • Feasibility study (types of feasibility): https://en.wikipedia.org/wiki/Feasibility_study
  • Decision table (limited/extended entry, rules): https://en.wikipedia.org/wiki/Decision_table
  • HIPO model (hierarchy plus input-process-output): https://en.wikipedia.org/wiki/HIPO
  • Nassi–Shneiderman diagram (structured-programming charts): https://en.wikipedia.org/wiki/Nassi%E2%80%93Shneiderman_diagram

Module 2 of 3 · 4 units · PDF pp.112–189

Module 2 — System Design (Input, Output, Files)

Page convention. The course pages carry printed numbers, but "p.N" always means the PDF page N of CIT212.pdf. Printed page = PDF page − 18 (PDF p.116 is printed p.98). Module 2 is PDF pp.112–189 (Unit 1 pp.112–125, Unit 2 pp.125–144, Unit 3 pp.144–163, Unit 4 pp.164–188). Section numbers (3.1, 3.5.4 …) and figure numbers are the course's own. Figures are numbered again from 1 in each unit.

Likely to be examined. The exam is computer-based, MCQ and fill-in-the-blank. The module gives these hooks:

  • Unit 1: objectives (p.113); self-assessment (p.119): name the five objectives of system design, list the constraints, list the processing techniques, define structured design; TMA (p.124): parts of the system design process, define "system specification", "prototype design".
  • Unit 2: objectives (p.126); self-assessment (p.137): list input devices, objectives of input design, methods of input verification and control; TMA (p.143): data dictionary, major concerns of CRT-input screen design, importance of colour in screen design.
  • Unit 3: objectives (p.145); self-assessment (p.153): fill-in blanks (a)–(e), considerations for output, general principles of output-report design; TMA (p.163): points in menu design, the three classes of forms, factors in form design, forms control.
  • Unit 4: objectives (p.165); self-assessment (p.178): file concept, types of files, the three methods of file organisation, basic factors for selecting media and organisation; TMA (p.188): advantages of a database, types of database structure, coding system, types of code.

The course guide (see Module 1 chapter) gives no list of examined topics, so treat every list in this module as a candidate for "name the …" and "which of these is NOT …" questions.

Jump to unit1234

Module 2 · Unit 1 of 4

Unit 1 — Structured System Design

What you must know

  • Name the five design objectives (practicality, efficiency, cost, flexibility, security) and the five designer constraints (hardware, software, budget, time scale, interface with other systems).
  • List the processing techniques (batch, real-time, online, a combination) and say why a combination is often ideal.
  • State what structured design is: a data-flow-based, top-down, modular methodology; define modularisation (decomposition) and the top-down design.
  • Describe the three development activities listed under structured design (database design, program design, system and program test preparation) and the difference between system testing and acceptance testing.
  • Explain system interface specification, audit considerations and the audit trail.
  • List the five parts of the design process (output, input, file, procedure, control design) and say why the process is iterative.
  • Define "system specifications" and "prototype", and list the conditions for and steps of prototyping.

Key ideas

Intuition. Feasibility study says what the system must do; design says how. The designer works under limits (hardware, money, time), aims at a few clear goals (objectives), breaks the system into small modules (structured design), decides outputs → inputs → files → procedures → controls (design process), writes it all down (system specification) and, if the requirements are unclear, builds a throw-away working model first (prototype).

Design starting point (p.112). Objectives from the feasibility study are the basis of design. Design is a creative process needing logical thinking (systematic moves toward the end product) and lateral thinking (ideas beyond the usual functions and equipment, so previous solutions are not forced onto new situations). Detailed study of the existing system may continue during design (p.113).

3.1 System design consideration (p.113). Design is "not a step adherence of clear procedures and guidelines"; much depends on the knowledge and experience of the designer. Problems come from constraints imposed by the user or from hardware/software limits.

3.1.1 Design objectives (p.114). The primary objective is to deliver the requirements specified in the feasibility report. Five objectives:

Objective Course wording
(a) Practicality The system must be stable and can be operated by people with average intelligence (the course prints "opened").
(b) Efficiency Involves accuracy, timeliness and comprehensiveness of the system output.
(c) Cost Aim for minimum cost subject to satisfying all the requirements.
(d) Flexibility Modifiable as user needs change, without extensive reconstructing or recreation of software; also portable to different computer systems.
(e) Security Covers hardware reliability, fall-back procedures, physical security of data, and provision for detection of fraud and abuse.

Logical vs physical design (p.114).

Logical design Physical design
Describes The structure and characteristics of outputs, inputs, files, databases and procedures The physical construction that follows logical design
Produces A description of what the system has Actual program software, files and a working system
Order First Second

3.1.2 Constraints (p.115). The designer normally works under:

Constraint Meaning
Hardware Existing hardware affects the design.
Software Available operating system, utilities, language etc. constrain the design.
Budget Affects the scope and depth of design.
Time scale The system may be needed by a fixed time (e.g. start of a financial year).
Interface with other systems Data may come from or go to another computerised system, so files must be compatible in format and run on a certain processing cycle.

3.1.3 Processing techniques (p.115). Batch processing, real-time processing, online processing, or a combination of all of them. A combination is often ideal in traditional data-processing applications: it increases throughput and lowers online response time. Most business applications accept 24-hour-old data, so voluminous updates can run after office hours in batch mode.

3.1.4 Operation (p.116, Figure 3.1). The data flow through a system, and the designer must specify the requirements of each area:

User raising input → Data preparation → Data validation → Processing → Output handling → Action on output

3.2 Design methodologies (p.116). New methodologies try to: improve productivity of analysts and programmers; improve documentation and later maintenance and enhancement; cut cost overruns and delays drastically; improve communication among user, analyst, designer and programmer; standardise the approach to analysis and design; simplify design by segmentation (six aims).

3.3 Structured design (pp.116–117).

  • A data-flow-based methodology. It starts with a system specification that identifies inputs and outputs and describes the functions; this is the basis for a graphic representation.
  • Next the modules and their relationships are defined in a structure chart, using a data dictionary and other structured tools.
  • Logical design goes top-down: general features (reports, inputs) first, then each in detail. The program is partitioned into small independent modules arranged in a hierarchy that approximates a model of the business area.
  • Subdividing a problem into smaller segments is called modularisation or decomposition (pp.116–117, 119).
  • A design is top-down if it is a hierarchy of modules, each with a single entry and a single exit subroutine (p.117). Advantages: critical interfaces are tested first; early (incomplete) versions resemble the real system; structuring provides control and improves morale; the procedural characteristics define the order of processing.

3.3.1 Major system design activities (pp.117–118). The course says the activities are "database design, implementation planning, system test preparation, system interface specification, and user documentation", then describes three:

Activity Content
(a) Database design Design of the physical database; a key issue is how the access paths are implemented.
(b) Program design Choice of programming language, flowcharting, coding and debugging before conversion; the operating system limits which languages can run.
(c) System and program test preparation Each aspect has its own test requirement; system testing is done after all programming and testing are complete; test cases cover every aspect of the proposed system, actual operations, user interface etc.; test requirements become part of the design specification (a prerequisite to implementation).

System testing vs acceptance testing (pp.117–118).

System testing Acceptance testing
Purpose Test every aspect of the system after programming is done Convince the user the proposed system meets the stated requirements
Technically Same kind of procedure "Technically similar to system testing"
Politically Done by the developers "Politically it is different": done in the presence of the user, audit representatives or the entire staff

3.3.2 System interface specification (p.118). Specifies for the user how information enters and leaves the system; the designer offers the user options. By the end of design, formats must be agreed so that machine-machine and human-machine protocols are defined before implementation. An operator's manual (how to install and operate, give input, access/update/retrieve information, display or print output and in what format) must be ready before implementation.

3.3.3 Audit consideration (p.118). A well-designed system has controls for proper, routine auditing; failure often comes from too little emphasis on data control. Standards of accuracy, consistency and maintainability must be specified to eliminate errors and control fraud. In a manual system, internal control depends on human judgement, personal care and division of labour; in a computer system fewer people are involved and software replaces human judgement in processing routines and error checks.

3.3.4 Audit control and documentation control (p.118). An audit trail is a routine designed to allow the analyst, user or auditor to verify a process or area of the new system. In a manual system it consists of journals, ledgers and other documents used to trace transactions. In a computer system, record content and format often make it difficult to trace a transaction completely, so the analyst must know basic auditing or work with an auditor. Documentation is the basis for review of internal controls by internal or independent auditors and a reference for maintenance. The main aim of auditing is to check that the controls built into the design ensure integrity; audit considerations must enter at an early stage so changes can be made in time (p.119).

3.4 Modularisation (p.119, Figure 3.2). The program is segmented into small independent modules in a top-down hierarchy with the details at the bottom; this minimises complexity. The top level shows the most important division of work, the lowest level the details.

3.5 Design process (pp.120–121). "Specifying how the system will work"; it is iterative and is based on what the system does as shown in the feasibility report. Five parts:

# Part What it does
(i) Output design Starting point: system requirements are converted into outputs.
(ii) Input design Decide what data are needed to produce the outputs; identify the source documents (revise or introduce new ones if needed).
(iii) File design Data captured are kept for short or long periods, stored in files in a logical manner; devise ways of storing and retrieving.
(iv) Procedure design How processing is performed: computer procedure (functions on the computer, which programs, in what sequence) and non-computer procedure (manual procedures for feeding input, receiving output).
(v) Control design Procedures to ensure correctness of processing, accuracy of data, timely output; the system works as planned.

The steps are inter-dependent and traversed many times ("two steps forward, one step backward"). Figure 3.3 shows the "ideal" progress (Study (analysis) → Design → Construction → Implementation, each stage once); Figure 3.4 shows the "reality" (loops back within and between stages). A change at one step has a knock-on effect on the others: if an output report is modified, input design, file design and control design may need to change. Aim at the most satisfactory and operable design, then improve gradually; trying to design a "perfect system" may leave you with no system at all (p.121).

3.6 System specifications (p.122). The result of the design process is the document called system specifications: all details of the proposed system; a blueprint for developing and implementing it; the primary documentation for the maintenance staff. Later it is divided into the system manual, user manual and operational manual. It may need modification after practical difficulties, bottlenecks or errors are found in later stages.

3.7 Prototype design (pp.122–123). A prototype is a working system developed to test ideas and assumptions about the new system; the first version of an information system, an original model, "a pilot or test model" designed to be changed. What is learned goes into a modified design, which may again be a prototype; repeat as often as necessary.

Prototypes are most useful when (p.122):

  • No system with the characteristics of the proposed one has yet been built by the developers.
  • The essential features are only partially known, and others cannot be identified even by careful analysis.
  • Experience of using the system will significantly add to the requirements.
  • Alternate versions will evolve through experience and refinement.
  • The users will participate in development.

Underlying principles (p.123): users can point to likes and dislikes in a working system more easily than they can describe them for a theoretical system; prototyping is interactive (course's word): it may start with a few functions and expand, or start with a "complete" set that expands or contracts.

Steps of the prototyping process (p.123): (1) identify the user's known information requirements and features; (2) develop a working prototype; (3) use it, noting needed enhancements and changes; (4) revise the prototype from user experience; (5) repeat as needed until satisfactory. It is not trial and error: user and analyst first discuss and agree the requirements; the system analyst is fully responsible for developing the prototype.

Code and worked examples

Example 1 — The data flow of Figure 3.1 (Course p.116; illustration of a library book issue added by us).

Text
User raising     Data           Data          Processing     Output        Action on
input        ->  preparation -> validation -> (update      -> handling -> output
(member asks     (fill issue    (member and   book status)   (due-date     (member takes
 for a book)      slip)          book exist?)                  slip)         book, returns it)

Course point: for each of the six areas, the designer must consider and specify the requirements.

Example 2 — Decomposition (Course Figure 3.2, p.119). The tree redrawn as text (the label on the top node is cut off in print; the unit's text calls it the inventory application):

Text
Inventory control
├── Stock status of books
│   ├── Issue of books
│   ├── Receipt of books
│   └── ... (etc.)
└── Status of journals
    ├── ...
    └── ...

Reading it: the top level shows the most important division of work; the lowest level shows the details; each box is one small module.

Example 3 — "A change at one step ripples to the others" (Course p.121; a worked trace, the new column is our illustration).

Step Change
Output design Report gets a new column, "Department".
Input design A source document and screen must now capture "Department".
File design The record needs a Department field with length and type.
Control design A validity check (code check) for Department is added.

Result: one change in output design forces changes in input, file and control design, so the design process is repeated, not done once.

Example 4 — Sorting statements into the right list (exam style).

Statement Belongs to
"System must work with the existing network" Constraint: hardware
"System must be live by the start of the financial year" Constraint: time scale
"Receives data from the payroll system, same file format needed" Constraint: interface with other systems
"Outputs must be accurate, timely and comprehensive" Objective: efficiency
"Fall-back procedures and detection of fraud" Objective: security
"Update large volumes after office hours, answer queries online" Processing technique: a combination
"Specify what runs on the computer, which programs and in what sequence" Design process: procedure design (computer procedure)

Example 5 — Prototyping loop (Course p.123), traced on a small illustrative case.

Round Step What happens
1 Identify requirements Registry staff know they need student lookup by roll number.
1 Develop prototype Analyst builds a lookup screen.
1 Use it Staff ask for lookup by name as well (new requirement).
2 Revise Name search added.
2 Use it Staff want a printout. The list of requirements grows again.
3 Repeat until satisfactory Staff accept it; the final requirements go into the system specification.

Common mistakes

  • Objectives vs constraints. Practicality, efficiency, cost, flexibility, security are objectives; hardware, software, budget, time scale, interface with other systems are constraints. "Cost" is an objective, "budget" a constraint.
  • Logical vs physical design. Logical describes the structure and characteristics (outputs, inputs, files, databases, procedures); physical produces actual software, files and a working system. Logical comes first.
  • Order of the design process. It starts from output, not input: output design → input design → file design → procedure design → control design. The input depends on what the output needs.
  • System testing vs acceptance testing. Technically similar; the difference is who is convinced and who is present (user, audit representatives, the entire staff).
  • Prototype vs final system. A prototype is designed to be changed and is a pilot or test model; it is not trial and error and it is not used only when requirements are fully known (the opposite).
  • System specification vs feasibility report. Specifications are the result of design (the blueprint, later split into system manual, user manual, operational manual); the feasibility report states the objectives design starts from.
  • Top-down vs bottom-up. The course's structured design is top-down; the hierarchy shows the most important division at the top and the details at the bottom.

Errata

Page Course says Correct Note
p.112, p.119 Contents entry "3.4 Modulation"; the heading on p.119 is "3.4 Modularization". Modularisation Same topic, different spelling in the contents; the exam word is modularisation/decomposition.
p.116 Heading "3.3 Structures Design" (contents p.112: "Structured Design") Structured Design Typo in the heading.
p.114 "provision for decision of fraud and abuse" (security) "detection of fraud and abuse" Typo; the meaning is detection.
p.114 Practicality: system "can be opened by people with average intelligence" "operated" Typo.
p.117 The structured-design activities are said to be five: "database design, implementation planning, system test preparation, system interface specification, and user documentation"; the sub-items are (a) Database design, (b) Program design, (c) System and program test preparation. The two lists do not match: "implementation planning" and "user documentation" are not sub-items, and "program design" is not in the opening list. Interface specification and documentation are covered in 3.3.2. For MCQs, learn both: the opening list of five and the headed three.
p.118 Section "3.3.4 Audit Control and Documentation Control" Not listed in the unit contents (p.112 lists 3.3.1–3.3.3 only). Content is examinable (audit trail, documentation).
p.121 "decisions made or changed at one step will have a 'triple effect' on other steps" The usual term is "ripple effect". Likely misprint. If an MCQ says "triple effect", it follows the course.
p.121, p.123 Calls the design process (and prototyping) "interactive" (the same unit says "iterative" on p.120). "iterative" in sense (repeated), as the text itself ("traversed many times", "repeat the steps") describes. Course's own version: use both words as printed.

Practice questions

  1. (Self-assessment i) Fill in: the five objectives of system design are practicality, efficiency, ______, flexibility and ______.
  2. (Self-assessment ii) Which of these is not one of the designer's constraints listed in the course? A. Hardware B. Budget C. Time scale D. User training

  3. (Self-assessment iii) Which of these is not one of the processing techniques the course lists? A. Batch processing B. Real-time processing C. Online processing D. Distributed processing

  4. (Self-assessment iv) Fill in: structured design is a ______ based methodology that partitions a program into small, independent ______.

  5. In the design process, which part is the starting point? A. Input design B. Output design C. File design D. Control design

  6. A design is top-down if it is a hierarchy of modules, each with a ______ entry and a ______ exit.

  7. Acceptance testing differs from system testing mainly in that: A. it uses different test data B. it is done before programming C. it is technically similar but conducted in the presence of the user and audit representatives D. it is only done by programmers

  8. (TMA 3) Prototypes are most useful when: A. the essential features of the system are fully known B. the essential features are only partially known and use will add to the requirements C. the system is a small payroll program D. the user does not want to take part

  9. (TMA 2) The document that results from the design process, a blueprint later split into the system manual, user manual and operational manual, is called the system ______.

  10. An audit trail is a routine designed to allow the analyst, user or auditor to: A. design the database B. speed up processing C. verify a process or an area in the new system D. print reports
Answersshow
  1. Cost; security. The five in 3.1.1 are practicality, efficiency, cost, flexibility, security.
  2. D. The five constraints are hardware, software, budget, time scale, interface with other systems.
  3. D. The course lists batch, real-time, online and a combination of them.
  4. Data flow; modules. p.116 ("a data flow based methodology"), p.116 (small, independent modules).
  5. B. The starting point of the design process is the proper knowledge of requirements converted into output.
  6. Single; single. p.117.
  7. C. p.117–118: technically similar, politically different.
  8. B. p.122: essential features only partially known; experience adds requirements.
  9. Specifications. p.122 (3.6). TMA 1 answer in full: output, input, file, procedure, control design.
  10. C. p.118 definition of audit trail.

Module 2 · Unit 2 of 4

Unit 2 — Input Design and Control

What you must know

  • Explain why input design matters and compare batch and online processing of transaction data.
  • Describe source documents and what an operator needs to know about each input field.
  • List the input media and devices (punch cards, key-to-diskette, MICR, mark-sensing, OCR, optical bar-code readers, CRT screens).
  • State the five objectives of input design and the five major concerns about input.
  • Apply the ten input verification and control methods to a given input error.
  • Say what a data dictionary stores and why the analyst uses it.
  • Apply the rules for CRT-input screen layout, and discuss ease of use, speed, menus, emphasis, colour use and selection, and editing/deleting through screens.

Key ideas

Intuition. Bad input gives bad output ("the input data are the lifeblood of a system"). Input is expensive and slow because people do it. So design input to be small, quick, simple and checked, and check it before it spoils the files.

Why input design (p.126). Input is one of the most expensive phases of the operation of a computer system, and many problems with a system can be traced to faulty input design. Input design features ensure reliability and correct reports, and decide whether the user can interact efficiently with the system.

3.1 Processing transaction data (pp.126–128). A transaction processing system keeps records of routine business activities, follows standard operating procedures and does not need complex decision-making; it needs accurate, detailed data and must process large volumes quickly. Examples: accounting, inventory, sales-order entry, banking, airline reservation. The cycle: enter data → process data → output.

Batch processing (3.1.1, p.127) Online processing (3.1.2, pp.127–128)
Idea Data collected for a period; the resulting "batch" is processed as a single job. Data processed instantly by the CPU; each request processed individually, answered within a few seconds.
Best when Most records in a large database must be processed at the same time. Selected records must be processed at any single time, or the user and computer must interact.
Examples Billing (telephone, gas, electric, cable TV), payroll (weekly/monthly). Bank teller terminal, airline reservation terminal.
Risks / extras An OLTP failure halts many activities (e.g. bank deposit system); remedies: fault-tolerant systems, disk mirroring (mirror image of critical disk data on two physically separate drives). OLTP is multi-user and multi-tasking. Tapes/optical storage serve as backup.

3.2 Elements of input data (p.128). Inaccurate input is the most common cause of errors in data processing, and input design can control operator errors. Goal: make data entry as easy, logical and error-free as possible. Operators need to know: (1) the allocated space for each field; (2) the field sequence, which must match the source document; (3) the format in which data fields are entered (e.g. date as mm/dd/yy). Approach: design the source documents that capture data, then select the media used to enter them.

3.2.2 Source documents (p.129). Source data are first captured on original paper, a source document (e.g. a cheque, encoded with MICR at the bank). Source documents start the processing cycle when entered (from punch cards, diskettes or the keyboard). Each source document is evaluated by: its continued use in the proposed system; the extent of modification; replacement by an alternative document. A source document should be logical and easy to understand: each area clearly identified, saying what to write and where. A date of birth can be written four ways (5 December 1994; Dec. 5 1994; 12/5/94; 5/12/94, European style) unless the form fixes MM/DD/YY.

3.3 Input media and devices (pp.129–130).

Medium or device Course description
(a) Punch cards 80 or 96 columns; data sequential and logical; keypunch copies from source documents, so document and card design are considered together.
(b) Key-to-diskette Modelled after the keypunch; a diskette replaces the card and stores up to "3250000" characters, "equivalent to 4050 punch cards" (see Errata); data kept in sequence and in batches.
(c) MICR Translates special magnetic-ink fonts on cheques into direct computer input.
(d) Mark-sensing readers Convert pencil marks in set places on a card into punched holes on the same card.
(e) OCR readers Like MICR but recognise pencil, ink or characters by their shape, not magnetic pattern; used in remote locations as free-standing preparation devices or direct input.
(f) Optical bar-code readers Detect combinations of marks that represent data; best known is the Universal Product Code (UPC) for retail items; speeds customer service and removes costly data-entry errors at point of sale.
(g) CRT screens For online data entry; normally 80 characters across and up to 24 lines.

Besides the medium, the analyst decides the method of input and the speed of capture. Processing may be batched (a group of records handled as a unit), online (records processed directly), sequential (sorted records) or random (unsorted). Example: magnetic tape suits batch sequential processing; diskettes suit online processing and random inquiries.

3.4 Input design objectives (pp.130–132). The design of input ties the information system to its users. Five objectives:

Objective Course explanation
3.4.1 Controlling the amount of data Entry depends on people, so labour cost is high; input is slow and the computer may sit idle; less input = lower cost and faster results.
3.4.2 Avoiding delay A delay caused by data preparation or entry is a bottleneck; avoid bottlenecks.
3.4.3 Avoiding errors in data Fewer data, fewer chances for error; design how data must be entered; detect errors with input validation techniques (checks and balances in the entry programs).
3.4.4 Avoiding extra steps When volume cannot be reduced (cheque runs, large retail), make the process efficient; one saved or added step is multiplied many times a day.
3.4.5 Keeping the process simple The best-designed system fits its users; avoid complexity when simple alternatives exist.

3.4.6 Major concerns regarding input (pp.132–135). Five questions: what input is needed? how and where is it created? how should the source documents be designed? what format for input records? what medium for recording input?

  • Input needed: determined by the output desired. Ask: what is already in the master file or database? what constant data can come from a control record? what must come from a transaction file? what can be held in tables? what can the program calculate?
  • How and where data are generated: enter data directly where it is created; remove manual keeping of data where possible. Course examples: cost accounting (data from factory, data-collection terminals); retail (bar-code readers, charge-card readers, POS terminal where the clerk keys only the quantity); electricity bills (the computer prints an input sheet by house number with meter number, last reading and customer type; the meter reader fills only the present reading); railway reservation (screen needs train code, travel class, date); Assam tea gardens (hand-held terminals in mobile vans record plucker code and weight; weekly payment reports are produced).
  • Source document and input format: designed together (p.134). Analyst works with the data-entry supervisor; documents should let recorders write early and fast; check boxes save time and reduce errors. Electricity-bill example: "MAIN" part (given to the customer) and "STUD" part (sent to the computer centre for the payment update). Field locations should be specified on the document; an identical layout for document and data-entry screen speeds operators. The input record's flow should match the source document. Factors: record length, field size, use of codes, relationship between source document and input record. A terminal screen should look like the source document.
  • Field size (p.134): determined by studying historical data, projecting future needs and "purchasing for growth" (p.135). Course example: 9,945 customers → allow five positions.
  • Input method (p.135): choose one needing minimal data conversion. Data usually go to a source document, then to the machine-processable medium; in process control, the medium-based programs directly control the target machines.

3.5 Input verification and control (pp.135–137). Wrong data are costly to correct (e.g. keying 100 instead of 10 sports cars). Methods:

# Method What it does (course example)
3.5.1 Key verification A second operator re-keys the data; the two versions are compared and mismatches are checked against the original documents. Used for punched cards, diskettes, magnetic tape.
3.5.2 Self-checking numbers (check digits) The computer rejects numbers that are transposed or have wrong digits; used for student roll numbers, account numbers, part numbers, invoice numbers.
3.5.3 Visually displaying an identifying characteristic Operator enters a part number; the terminal displays the description, and the operator confirms it visually.
3.5.4 Hash totals Add numbers to get a meaningless total (e.g. all quantities); compare with the original total after entry and processing; agreement shows all quantities were entered and all records processed.
3.5.5 Checking between a range of numbers Order numbers for the day must fall between 4999 (last of the previous day) and 6001 (next order number to be used); otherwise an error message.
3.5.6 Reasonableness test Based on past history, e.g. no employee has more than 20 hours' overtime; for date of birth, day ≤ 31, month ≤ 12, year not beyond the current year or current year minus minimum age.
3.5.7 Verification and codes Valid payroll status codes only H (hourly), S (salaried), T (trainee), P (part-time).
3.5.8 Verification of data type Numeric fields hold only numbers; alphabetic fields only letters.
3.5.9 Verification that certain combinations of data exist Students coded W (work-study) or V (non-work-study); only account numbers 2155 and 2156 are valid for a W student.
3.5.10 Sequence check If source documents are serial and in order, input records must be in ascending or descending numerical sequence.

3.6 Data dictionaries (p.137). A data dictionary stores descriptions of data items and structures and system processes. The analyst uses it during design (data length, alternate names (aliases), use in particular processes). It holds validation information that helps specify controls on accepting data, and definitions of data flows, data stores and processes. It can be manual or automated; automated ones produce listings of data elements, data structures and processes automatically, do cross-reference checking and error detection, and are becoming the norm. (More in Module 1, Unit 4.)

3.7 Laying out a terminal screen (p.138). Software lets a programmer key in a screen format, name it and store it in a file, then call it by name in a program. 3.7.1 CRT-input screens tell the user what to do next, briefly and clearly; menus present options; fields are marked to show length and where to enter data; valid entries must be identified for the programmers.

3.7.2 Basic rules for CRT-input display screens (pp.138–139):

  • Clear the entire screen between formats.
  • Format for easy reading: no clutter; directions or error messages always in the same place; space between items.
  • Prevent scrolling: show one screen at a time and move on when the operator presses a key.
  • Don't overuse colour.
  • Be consistent (e.g. all instructions at the bottom).
  • Use simple conventions (an operator enters '1' or 'Y' for a positive answer).
  • Make sure all directions are clearly stated.
  • Test all screens with someone unfamiliar with the program, who enters data without the help option, before release.

3.8 Major concerns of CRT-input screen design (pp.139–142). Seven: ease of use, improved processing speed, menu-driven screens, emphasising information, colour use, colour selection, editing through display screens.

Concern Course content
3.8.1 Ease of use Fill-in-the-blank approach: all data elements clearly labelled, space for entry; aesthetically designed; meaningful descriptions and error messages; avoid symbols and over-abbreviation.
3.8.2 Improved processing speed Abbreviated responses (Y for Yes); entry order consistent with the business transaction (no unnecessary tabbing); protected (cannot be changed) vs unprotected (can be changed) fields; program function keys can trigger a transaction with one press; terminals with application-specific features (number pad, POS terminals).
3.8.3 Menu-driven screens A menu shows what functions can be performed and how to select them; top-down menus lead the user through a series of decisions; uncluttered compared with narrative instructions.
3.8.4 Emphasis Blinking, underlining, increased/reduced light intensity, inverse video (black letters on light screen) for errors and messages; also confirm that data were accepted and processing has begun.
3.8.5 Colour use Gives structure and meaning; ties related items together. Four primary uses: (1) identify valid operations the user can carry out; (2) tie related data together; (3) highlight information about organisation performance; (4) communicate messages about system performance.
3.8.6 Colour selection Excess or improper colour loses meaning; no more than four colours can be quickly recognised; most useful: red, green, yellow, blue; effective contrast pairs: red/green, yellow/blue, white/blue; colour only when there is a reason.
3.8.7 Editing Editing = changes to stored records or records submitted but not yet stored; includes deletion. Two delete methods: (1) key that deletes the current record on the screen; (2) user enters the record key (e.g. item number) and then presses a delete key. Both common; the first is preferred because the user views the record before deleting.

Order of legibility (p.142) (printing colour on background colour; 1 is most legible):

Order Printing Background Order Printing Background
1 Black Yellow 8 White Red
2 Green White 9 White Green
3 Red White 10 White Black
4 Blue White 11 Red Yellow
5 White Blue 12 Green Red
6 Black White 13 Red Green
7 Yellow Black

Code and worked examples

Example 1 — Batch vs online (Course p.127; easy). A telephone company bills 2 million customers once a month: all records are processed together → batch. A bank customer asks for a balance at an ATM and gets it in seconds, one request at a time → online. A bank deposit system that fails stops all the branches, which is why disk mirroring keeps the same data on two separate drives.

Example 2 — Range check (Course p.136). Rule: today's order numbers lie between 4999 (last yesterday) and 6001 (next to be issued), so valid numbers are 5000–6000.

Order no. entered Test Result
5000 4999 < 5000 < 6001 accept
5345 in range accept
4999 yesterday's last order error message
6001 not yet issued error message
7210 outside range error message

Example 3 — Hash total (Course p.136). Five quantities on the batch header: 12, 7, 30, 5, 18.

  1. Hash total before entry = 12 + 7 + 30 + 5 + 18 = 72 (a meaningless total).
  2. The operator keys 12, 7, 30, 5, 17 (a slip). The computer total = 71.
  3. 71 ≠ 72, so the batch is rejected for checking against the originals.

Limit (our remark, not in the course): two errors that cancel (e.g. 12 and 7 swapped between two records) leave the total unchanged, so a hash total alone can miss them. A transposed pair inside one number (12 keyed as 21) changes the total (81 instead of 72), so it is caught.

Example 4 — Field size (Course p.134). 9,945 customers now; the field holds the customer number.

Positions Largest number held Verdict
4 9,999 9,945 + 54 = 9,999 still fits; the 55th new customer (10,000) does not fit.
5 99,999 Room for 90,054 more customers.

Course answer: allow five positions ("purchasing for growth").

Example 5 — A self-checking number (the course gives no formula; this modulus-11 sketch is our illustration). Weights 5, 4, 3, 2 for the account number 1234:

  1. 1×5 + 2×4 + 3×3 + 4×2 = 5 + 8 + 9 + 8 = 30.
  2. 30 mod 11 = 8, check digit = 11 − 8 = 3, so the stored number is 12343.
  3. A clerk transposes the first two digits and keys 21343: 2×5 + 1×4 + 3×3 + 4×2 = 31; 31 mod 11 = 9; expected check digit = 2 ≠ 3, so the number is rejected.

Example 6 — Menu dialogue (Course p.140), drawn as screens.

Text
SALES SYSTEM - MAIN MENU            EDIT MENU                    EDIT BUDGET
1. ...                              1. EDIT BUDGET   <-- press   (screen to edit the
2. ...                              2. ...                        selected budget)
3. EDIT MENU      <-- press 3       3. ...

The user presses 3 (EDIT MENU), then 1 (EDIT BUDGET, printed "I" in the course; see Errata). The screens stay uncluttered, unlike the long narrative sentences that would otherwise describe each step.

Common mistakes

  • Verification vs validation. In this unit the verification and control list (3.5) mixes both: key verification (re-keying by a second operator) checks that data were copied correctly; range, reasonableness, code, type, combination and sequence checks test the data's plausibility. Do not swap them in an MCQ.
  • Hash total vs check digit. A hash total is a meaningless sum of a field over a batch, compared before and after entry; a self-checking number (check digit) is built into one number to catch transposed or wrong digits.
  • Range check vs reasonableness test. Range check: order numbers between yesterday's last and tomorrow's first. Reasonableness: based on past history/policy (overtime not over 20 hours; date of birth day ≤ 31, month ≤ 12).
  • Sequence check vs range check. Sequence check: the records are in ascending/descending order. Range check: each value lies between two limits.
  • Batch vs online. Batch = collected over a period, one job; online = instant, each request on its own. Billing and payroll are the course's batch examples; a bank teller terminal and airline reservations are online.
  • OCR vs MICR. MICR reads magnetic-ink fonts (cheques); OCR recognises the shape of pencil/ink characters. Mark-sensing converts pencil marks to punched holes.
  • Protected vs unprotected fields. Protected = cannot or should not be changed; unprotected = data that can be changed.
  • Colour numbers. Four colours at most; the four most useful are red, green, yellow, blue; contrast pairs are red/green, yellow/blue, white/blue.
  • Which delete method is preferred. The one that deletes the current record on the screen (the user sees it first), not the one that needs the record key.

Errata

Page Course says Correct Note
p.125 Unit heading "INPUT DESIGN AND CONTROL"; module contents (p.112) says "Input Designs and Control" Same unit Wording only.
p.129 A diskette "stores up to 3250000 characters of data – equivalent to the data stored in 4050 punch cards" 3,250,000 ÷ 80 columns = 40,625 cards (÷ 96 = 33,854), not 4,050. 325,000 ÷ 80 ≈ 4,062, so the figure was probably 325,000. Numbers are inconsistent as printed. If an exam asks, give the printed numbers.
p.134 If the five positions are not reserved, "as soon as 54 more records are added, the field will not be large enough" (9,945 customers) A four-position field holds up to 9,999 = 9,945 + 54; the 55th record overflows it. Off by one. The lesson (allow five positions) is unchanged.
p.139 Operator enters "a 'I' or 'Y' for a positive response"; p.140: user "depresses 'I' for the EDIT BUDGET option" Likely the digit "1" (the same dialogue uses "3" for a menu choice) Printed as a capital I. Likely misprint; verified only from context.

Practice questions

  1. (Self-assessment ii) Fill in: the five objectives of input design are controlling the amount of data, avoiding ______, avoiding ______ in data, avoiding extra steps and keeping the process ______.
  2. (Self-assessment i) Which of these is not an input medium or device named in the course? A. Punch cards B. MICR C. Optical bar-code readers D. Laser printer

  3. Telephone billing and payroll are course examples of: A. online processing B. batch processing C. real-time control D. disk mirroring

  4. A second operator re-keys the data and the two versions are compared. This is: A. hash total B. key verification C. reasonableness test D. sequence check

  5. Quantities are added to give a meaningless total that is compared with the original total after processing. This is a ______ total.

  6. A payroll status field accepts only H, S, T or P. Which check is this? A. Range check B. Verification of codes C. Hash total D. Visual display

  7. (TMA 1) A data dictionary stores: A. only the data in the master file B. descriptions of data items, structures, data flows, data stores and processes C. only screen layouts D. only user passwords

  8. (TMA 3) Colour has four primary uses on a screen. Which is not one of them? A. Identify valid operations B. Tie related data together C. Reduce the cost of the monitor D. Communicate messages about system performance

  9. How many colours can users quickly recognise, according to the course? ______ (at most).

  10. Of the two ways to delete a record through a screen, the preferred one is to: A. enter the record key then press delete B. press a key that deletes the current record shown on the screen C. delete the whole file D. wait for a batch run
Answersshow
  1. Delay; errors; simple. pp.130–132 (3.4.1–3.4.5; the first objective is "controlling amount of data").
  2. D. Section 3.3 lists punch cards, key-to-diskette, MICR, mark-sensing readers, OCR, bar-code readers, CRT screens; a laser printer is an output device (Unit 3).
  3. B. p.127: billing and payroll are batch.
  4. B. 3.5.1 key verification.
  5. Hash. 3.5.4.
  6. B. 3.5.7 verification and codes.
  7. B. 3.6, p.137.
  8. C. The four uses (3.8.5, p.141) are valid operations, tying related data, highlighting performance, communicating system messages.
  9. Four. p.141 ("No more than four colours").
  10. B. p.142: preferred because the user views the record before deleting.

Module 2 · Unit 3 of 4

Unit 3 — Output System Design

What you must know

  • Give the three reasons for computer output and classify outputs as application (three types) or operating output.
  • Name the output devices (printers, VDUs, COM, audio output) and say when printers are used.
  • List the eight output design considerations (user and purpose, content, format, frequency and timing, volume, sequence, quality, type of stationery).
  • Label the parts of a report (report heading, page heading, control heading, detail line, control footing, final control footing, page footing, report footing) and give the principles of report design, including the print chart.
  • Design screen output and menus (four menu design points).
  • Define a form, classify forms (action, memory, report), list the form design factors and explain forms control.
  • State the benefits and the two kinds of computer graphics (presentation, decision support).

Key ideas

Intuition. Output is what the user sees, so the system is judged by it. Decide who needs it, what it contains, how often, on what, and how big; draw the layout on a chart before programming; for screens use menus; for paper use well-designed forms; for summaries use graphics.

Why output is needed (pp.144–145). Three reasons: (i) to communicate to the persons concerned; (ii) for re-input to the computer, to be connected with other data and processed further; (iii) for permanent storage.

3.1 Types of output (pp.145–146). Outputs vary in content, stationery, frequency and timing; consider who will use the output and why. Two categories:

Category Meaning Examples (course)
Application output (3.1.1) Outputs wanted from the system to meet its objectives. Three types: (i) a basis for decision-making (management); (ii) a requirement to meet a functional objective; (iii) statutory outputs, required by law. (ii) invoices, excise gate pass, purchase orders; (iii) 'C' forms, '3A' and '6A' forms for provident fund, income tax certificates.
Operating output (3.1.2) Generated for the E.D.P. (electronic data processing) staff to show how the system operates; not for users. System logs, error messages, status indicators.

3.2 Output devices (pp.146–147). The most important: printers, video display units (VDUs), computer output microfilm (COM); also audio output.

  • Printers are used when (i) large volume of output is required; (ii) output goes to various persons inside/outside the organisation; (iii) batch processing is used. They give permanent visual records, can print on ordinary paper or on special forms (dispatch notes, invoices, packing slips), at "150–2500 lines per minute, each line as many as 150 characters".
Impact printers Non-impact printers
How A print head strikes a print ribbon which prints a character on paper. No mechanical print head; a jet of ink or a laser beam prints the character.
Cost Cheaper "A little bit costlier"
Types / notes Line printers (one whole line at a time; about 1000–1200 lines per minute; expensive; large volume). Character printers (one character at a time; cheaper, slow, about 80–90 characters per minute; low-volume use such as word processing). Laser printers are widely used in Desktop Publishing (DTP) for excellent print quality.
  • VDUs use cathode ray tubes (CRTs); they have a keyboard for entering data; 24 lines of 80 characters; widely used in online systems to show results and answers to queries. Colour units show colours. Displayed information may be plotted on a plotter or recorded on a diskette or video tape recorder.
  • COM records computer output as microscopic images on roll or sheet film; the images are the same as would be printed on paper; the recording reduces characters 24, 42 or 48 times smaller; recorded on 16 mm, 35 mm microfilm or 105 mm microfiche. The data come online from the computer or off-line from magnetic tape. In most recorders the data are displayed on a CRT, a camera photographs it onto film, and the film is processed.
  • Audio output announces results on a loudspeaker; the bit pattern of a word is stored in an audio output buffer storage and converted to speech by a speech synthesiser; most common use: answering enquiries such as a telephone number or a savings account balance; speech is not yet natural.

3.3 Output design considerations (pp.148–149).

Consideration Meaning
Type of user and purpose Different users need different outputs: exception reports (e.g. sales fall below a level), summary reports (sales quantity and value per region), or details (list of invoices for a period); statutory reports follow the law, little flexibility.
Content The data to include, related to the purpose.
Format Arrangement of data, size of paper, titles, headlines, colour of paper.
Frequency and timing How often (daily, weekly, monthly, annually) and when (after annual closing of accounts, end of fiscal year, before the last day of every month).
Volume Sheer volume deters use; creates handling, filing and printing-time problems.
Sequence Usefulness depends on the sequence (e.g. pay-slips printed department-wise help payment).
Quality Content, appearance and accuracy; extra attention for external users (get-up, paper quality).
Type of stationery Blank stationery, or pre-printed stationery when most of the content is constant (invoices, pay-slips); advantages: saves computer time, attractive and convenient; disadvantage: costlier.

3.4 Design of output reports (pp.149–152). A report normally has: a report heading (first page only); a page heading and sub-heading (top of each page); sets of records with common features grouped as a control group, with a control heading. Terminology (Table 1, the NOUN roll list):

Term Meaning In the course's roll list
Report heading Appears once for the report NATIONAL OPEN UNIVERSITY OF NIGERIA – ROLL LIST OF MIT STUDENTS
Page heading Top of each page ROLL LIST OF STUDENTS FOR DIFFERENT COURSES OF SEM. 1/2006-7
Control heading and sub-headings Start of each control group LIST OF STUDENTS IN PASCAL / ROLL NO. NAME OF STUDENT
Detail line One line of detail below the heading 1001 Iyioluwa Adegboye Jegede
Control footing Abstract of a control group at its end TOTAL NO. OF STUDENTS IN PASCAL = 52 (the explanation on p.151 prints 50; see Errata)
Final control footing Overall abstract Total No. of students in semester 1/2006-7 = 570
Page footing Information at the end of each page
Report footing Information at the end of the report END OF NOUN ROLL LIST FOR SEMESTER 1/2006-7

Print chart (p.152). The analyst decides the report layout on a print chart: it depicts the number of columns and lines; the number of columns normally equals that of the printer; headings are entered on the chart to see how the output will look. The key concern is a layout that is easy to read. The chart is then handed to a programmer using a report generator (e.g. COBOL's report generation program). The convention in the course's Figure 1: 9 for digits, X for characters, B for blank, Z for digits or leading zeros, $ for currency symbols. Sample (Figure 1): xx99999 xxxxxxxxxxx 99B99B99 99,999.99, and a final line **PAY-TOTALS**999,999.99.

General principles of report design (p.152):

  • It must be readable left to right and top to bottom.
  • The most important item, such as the key field, is easy to find.
  • All pages have a heading and a page number; the date of preparation is printed too.
  • Avoid too many details.
  • Use control footing to abstract information about groups of detail lines.
  • Define page and report footings properly.

Government reports (e.g. to the tax department) follow the agency's format strictly. When many forms are printed (marks records, share certificates, dividend warrants), pre-printed stationery carries the constant text and the program prints only the variable data (p.152–153).

3.5 Screen output (p.154). Most print-design principles apply, with differences: a screen has normally 80 columns and 24 rows; the user must actively take part (online), so proper instructions for retrieving information are needed; the screen has a heading and field headings; instructions at the bottom guide the user to the next screen or to exit; the system must recover if a wrong key is pressed (e.g. E in the MIT screen) by giving a message and showing the alternatives again.

3.6 Menu design (pp.155–157). Four points:

Point Rule
(i) Hierarchical Menus in a structured hierarchy from higher-level to lower-level functions (order processing: Data entry, File maintenance, Processing, Inquiry, MIS reports, Exit); the next level appears depending on the choice.
(ii) Termination When a program under a menu ends, show again the menu from which it was started; termination from one level leads to the menu of the next level; each menu should have an option to terminate completely or return to the highest level.
(iii) Skipping menu Where functions are restricted to particular persons, allow a jump to the screen directly depending on the user code.
(iv) Security control Not everyone may do every function, so menus may be segmented by function to control access (e.g. a clerk in the Income Tax Dept. only enters returns).

3.7 Form design and control (pp.157–159).

  • Form (3.7.2): the physical carrier of data and of information; it can constitute authority for action: a purchase order says BUY, a customer's order says SHIP, a paycheque says PAY TO THE ORDER OF. Each form is a request for action; poorly designed forms are costly. Most forms are designed by poorly trained people.
  • Classification (3.7.3, Fig. 2):
Class What it does Characteristics Examples
Action Requests the user to do something Orders, instructs, authorises; achieves results; goes from one place (person) to another Application form, purchase order, sales slip, shop order, time card
Memory Record of historical data that remains in a file Represents historical data; used for reference; stationary, stays in one place, usually a file; serves as control on certain details Bond register, inventory record, journal sheet, purchase record, stock ledger
Report Guides supervisors and administrators; data on a project or job Summary picture of a project; details about a job or things needing attention; used by a manager with authority to effect change; basis for decision-making Balance sheet, operating statement, profit and loss statement, sales analysis, trial balance
  • Factors in form design (3.7.4), seven: (i) size and shape convenient for handling, filling, sorting; (ii) material in logical order; (iii) a title that clearly identifies the purpose, with labelled columns and rows; (iv) precise content, adequate and compact space, pre-printed entries; (v) special features like security and control; (vi) emphasis by shading columns, heavy lines etc., and detail arranged for operations such as copying or checking; (vii) covers the specific needs of the purpose.
  • Forms control (3.7.5): forms multiply and cost money, so large organisations set up a formal forms control program. Its first objective is to establish standards (different departments using different forms for the same task is an unnecessary expense). The forms control specialist eliminates redundancy to reduce clerical cost and reduces the number of copies (routing one copy through several departments is probably the best way). Forms are titled, numbered and carry the date of the most recent revision; form numbers are organised so all forms of a system can be located. A form is designed by an analyst with the users; at a re-order level the specialist handles re-ordering and revision, routes the form to users for comment, and orders the most economical list; for routine office forms, a re-order of one year's supply is normal.

3.8 Computer graphics (pp.160–162). Pages of text are tedious; summary data are best shown graphically. Use has grown with end-user computing, DBMS, spreadsheets, graphics software and high-resolution devices. Benefits (six): more effective conversion of data into information; easier recognition of relationships and trends; quick decisions; better presentation of output; ability to focus attention on important issues; attractive format that gets attention. Behind them: the mind absorbs information faster from a picture than from words or numbers. Software divides into two categories:

Presentation graphics (3.8.1) Decision support (analytical) graphics (3.8.2)
Purpose Communicate ideas to people unfamiliar with a situation or needing a simple overview (insurance comparison, market share change, next-quarter workload) Understand patterns, trends, relationships in data; learn something about the data
Quality High-quality illustration, range of colours, fonts, reduce/enlarge, 35 mm slides or transparencies, text mixed with graphics Quality of the illustration matters less than its ability to support problem-solving; colour and special graphical effects usually needed
Data source Several databases, non-computer sources, outside sources Spreadsheets, local databases or the firm's central database (the system must reach the data with minimum user involvement)

Pie chart example (p.161): the user types the title, then for each slice: label, value, whether it is "exploded", colour, design code (texture); the software does the rest. Graphics hardware (3.8.3): graphics terminals, graphics boards, graphics printers, interface devices; software gives fonts, font sizes, colours, moving images, graphics inside text, interface devices; packages offer line, bar and pie charts, some in three dimensions, several charts on one plot.

Code and worked examples

Example 1 — Labelling a report (Course Table 1, pp.149–151).

Text
NATIONAL OPEN UNIVERSITY OF NIGERIA - ROLL LIST OF MIT STUDENTS     <- report heading (once)
ROLL LIST OF STUDENTS FOR DIFFERENT COURSES OF SEM. 1/2006-7        <- page heading (every page)

LIST OF STUDENTS IN PASCAL                                          <- control heading
ROLL NO   NAME OF STUDENT                                           <- sub-heading
1001      Iyioluwa Adegboye Jegede                                  <- detail line
...
1052      Akpa Godson
TOTAL NO. OF STUDENTS IN PASCAL = 52                                <- control footing

LIST OF STUDENTS IN 'C' LANGUAGE                                    <- next control heading
...
TOTAL NUMBER OF STUDENTS IN 'C' LANGUAGE = 20                       <- control footing
TOTAL NO. OF STUDENTS IN SEMESTER 1/2006-7 = 570                    <- final control footing
END OF NOUN ROLL LIST FOR SEMESTER 1/2006-7                         <- report footing

Check: roll numbers 1001–1052 are 1052 − 1001 + 1 = 52 students (the course prints 52 in the table and 50 in its explanation); 1070–1089 are 20. A page footing (information at the end of each page) is not shown in the sample.

Example 2 — Reading a print chart (Course Figure 1, p.153). The pay report line xx99999 xxxxxxxxxxx 99B99B99 99,999.99 with the course conventions (X = character, 9 = digit, B = blank).

Field Picture Meaning Sample value printed
Employee no. xx99999 2 characters + 5 digits (7 columns) AB12345
Name xxxxxxxxxxx 11 characters JOHN OKORO
Date 99B99B99 day, blank, month, blank, year (8 columns) 05 11 26
Total pay 99,999.99 5 integer digits, point, 2 decimals (9 columns) 04,250.50

The last line shows the report total **PAY-TOTALS**999,999.99 (one more integer digit than a single pay amount). With a Z picture (ZZ,ZZ9.99, standard COBOL) leading zeros would be suppressed and show as 4,250.50; the course only says "Z for digits or leading zeros".

Example 3 — Menu path (Course pp.155–156).

Text
ORDER PROCESSING                    ORDER PROCESSING             FILE MAINTENANCE
1. DATA ENTRY                       (FILE MAINTENANCE)           (CUSTOMER MASTER)
2. FILE MAINTENANCE  <-- select 2   1. CUSTOMER MASTER <-- 1     1. ADD A RECORD
3. PROCESSING                       2. PRODUCT MASTER            2. MODIFY
4. INQUIRY                          3. DISCOUNT MASTER           3. DELETE
5. MIS REPORTS                      4. EXIT                      4. EXIT
6. EXIT                             WHICH ONE?                   WHICH ONE?
WHICH ONE?

Hierarchical: each choice opens the next lower level. Termination: after ADD A RECORD finishes, the Customer Master menu is shown again; EXIT goes up one level. Skipping: a user with a user code for customer maintenance may jump straight to the third screen. Security: a clerk may see only the options their user code allows.

Example 4 — Classifying outputs (exam style; Course 3.1).

Output Class
Sales trend summary for the managing director Application: basis for decision-making
Customer invoice Application: functional objective
Income tax certificate Application: statutory
System log of a program failure Operating output
Error message "invalid code" for the operator Operating output

Example 5 — Which consideration? (Course 3.3).

Situation Output design consideration
Report sent only when sales fall below a set level Type of user and purpose (exception report)
Pay-slips printed department by department Sequence
Monthly report printed after the last day of each month Frequency and timing
A 400-page printout nobody reads Volume
Invoices printed on pre-printed forms Type of stationery (saves computer time; costlier)
Report for outside customers on good paper Quality

Example 6 — Classifying forms (Course Fig. 2, p.158).

Form Class Why
Purchase order Action Says BUY; goes from one person to another
Stock ledger Memory Historical data kept in a file for reference
Profit and loss statement Report Summary used by management for decisions
Time card Action Orders/authorises, achieves results
Journal sheet Memory Historical record

Common mistakes

  • Application vs operating output. Application output is for users (decision, functional, statutory); operating output is for the EDP staff (logs, error messages, status indicators). An error message shown on the operator's console is operating output.
  • Control footing vs final control footing vs report footing. Control footing summarises one control group (total for Pascal); final control footing is the overall total (semester total 570); report footing is the closing line of the report ("END OF …"). Page footing is at the end of each page.
  • Report heading vs page heading. The report heading appears once (first page); the page heading is at the top of every page.
  • Impact vs non-impact printers. Impact printers (line, character) strike a ribbon; laser and ink-jet are non-impact (no print head) and costlier; laser is widely used in DTP. A line printer prints a whole line at a time; a character printer one character at a time.
  • Action vs memory vs report forms. Action = asks someone to do something (purchase order); memory = historical record kept in a file (stock ledger); report = summary for supervisors/managers (balance sheet).
  • Presentation vs decision support graphics. Presentation = communicate to others, high quality, slides; decision support = learn patterns and trends from the data, quality matters less, data from spreadsheets or central database.
  • Print chart vs report layout. The print chart is the grid (columns × lines) on which the analyst places headings and fields before programming.
  • Forms control vs form design. Design = what goes on one form; control = standards, numbering, fewer copies and forms, and ordering/re-ordering.

Errata

Page Course says Correct Note
pp.150–151 Table 1 prints "TOTAL NO. OF STUDENTS IN PASCAL = 52"; the explanation (v) prints "= 50". 52 (roll nos. 1001–1052). The two prints disagree; the table is consistent with its own numbers.
pp.150–151 Table 1 ends "END OF IGNOU ROLL LIST …"; the explanation says "END OF NOUN ROLL LIST …". NOUN (this course's university). Leftover from the IGNOU source text.
p.151 (iii) "The page headings and sub headings are as follows: LIST OF STUDENTS IN PASCAL / ROLL NO. NAME OF STUDENT" These are the control heading and sub-headings (the next lines say "Another control heading"). Page heading is the line above (ROLL LIST OF STUDENTS FOR DIFFERENT COURSES …).
p.152 "Z for digits or leading zeros" In standard COBOL, Z shows a digit but replaces leading zeros by spaces; 9 shows the digit including zeros. Verified by COBOL picture rules (outside the course). Use the course wording in the exam.
p.153 Self-assessment (e): "…… is an output technique that records output from a computer as microscopic images on roll or sheet film." The blank is missing; the answer is computer output microfilm (COM). Printing slip.
p.158 Fig. 2 lists "Trail balance" as a report form. Trial balance Typo.

Practice questions

  1. (Self-assessment i-a, b) Fill in: outputs take different forms: reports, ______ and printed forms; they fall into two categories, ______ output and ______ output.
  2. (Self-assessment i-c, d) The two types of printers are ______ and ______; ______ printers are widely used in Desktop Publishing.
  3. (Self-assessment i-e) The output technique that records output as microscopic images on roll or sheet film is ______.
  4. System logs, error messages and status indicators are examples of: A. statutory output B. decision-making output C. operating output D. functional output

  5. Pay-slips are printed department by department so that payment is easier. Which output consideration is this? A. Volume B. Sequence C. Format D. Frequency and timing

  6. In a report, the line "TOTAL NO. OF STUDENTS IN SEMESTER 1/2006-7 = 570" is the: A. report heading B. control heading C. final control footing D. page footing

  7. (TMA 2) The three primary classes of forms are ______, ______ and ______ forms.

  8. A stock ledger, which keeps historical data in a file for reference, is a: A. report form B. action form C. memory form D. statutory form

  9. (TMA 1) Which is not one of the four points to remember in menu design? A. Hierarchical B. Termination C. Colour matching D. Security control

  10. Graphics used to understand patterns, trends and relationships in data are called ______ graphics (also analytical graphics).

Answersshow
  1. Displays on screen; application; operating. 3.1: "reports, displays on screen, printed forms etc."; two categories application and operating.
  2. Impact; non-impact; laser. 3.2.
  3. Computer Output Microfilm (COM). 3.2.
  4. C. 3.1.2: for the EDP staff.
  5. B. 3.3 sequence (pay-slips department-wise).
  6. C. Control footings are the Pascal and C language totals; the whole-report total is the final control footing.
  7. Action, memory, report. 3.7.3.
  8. C. Memory forms keep historical data for reference (Fig. 2).
  9. C. The four are hierarchical, termination, skipping menu, security control.
  10. Decision support. 3.8.2.

Module 2 · Unit 4 of 4

Unit 4 — File and Database Design

What you must know

  • Describe the file hierarchy (data item, record, file, database) and the key field.
  • Select storage media and explain brought-forward/carried-forward files and generations.
  • Name the eight types of files (master, transaction, table, report, backup, archival, dump, library) and tell master, transaction, backup and archival apart.
  • Compare sequential, direct (random) and indexed file organisation, with advantages and disadvantages; use the division-by-prime hashing example.
  • List the six factors for file design (method of updating, size, inquiry capability, activity ratio, volatility, response time) and apply the 30 % activity-ratio rule.
  • Define a database, its objectives and advantages, DBMS, schema and sub-schema, and the three models (hierarchical, network, relational).
  • Define a code and describe the seven types of codes.

Key ideas

Intuition. After output and input comes storage: records must be kept so they can be found and updated quickly. Choose the medium (tape vs disk) and the organisation (sequential, direct, indexed) from how the file is used (how often, how fast, how large). Then move from separate files to a shared database to cut redundancy. Finally, codes shorten input and cut errors.

Heading note. The unit is printed "UNIT 4 INPUT DESIGN AND CONTROL" (p.164), a copy of Unit 2's title; its real title from the module contents (p.112) is File and Database Design (see Errata).

Introduction (p.165). After input and output, the designer organises data around user requirements. The analyst designs files and selects their contents, choosing among options. File organisation may be sequential, index sequential, inverted list or random (the intro lists four; section 3.3 treats three). An integrated approach to file design is the database: information handled as an integrated whole, with minimum redundancy and better performance. Objectives of a database: accuracy and integrity, recovery from failure, privacy and security of data.

3.1 Selecting data storage media (pp.165–166). The analyst is interested in bulk storage using magnetic media: magnetic tape and exchangeable magnetic disk. Magnetic drums and fixed discs are for special-purpose files; magnetic card devices have large capacity but are slow and not very reliable; punched cards and paper tape are obsolete for bulk storage.

  • Magnetic tape: reels of 0.5-inch plastic tape, 2400 ft, magnetisable coating on one side; the tape moves from spool to spool; you cannot read and immediately write on the same reel. Updating a file therefore means reading the existing file from one reel and writing a new version on another. For a master file updated from time to time, the file being read is the brought-forward file and the new one is the carried-forward file. Different versions are generations, useful for recovery; three generations (grandfather, father, son) are commonly kept before the oldest is reused.
  • Data density: bits per inch (bpi); rows of seven or nine spots across the tape (7- or 9-track tape); each row (frame) is one character (byte); data are recorded in blocks of specified size.

3.1.1 File concepts (pp.166–167). The hierarchy: data item → record → file → database.

Term Course definition
Data item (field) The basic, individual element of data, identified by a name and assigned a value (e.g. Employee Name = PANKAJ, Employee Identification Number = 1775).
Record A collection of related data items (e.g. an employee record with eight fields: name, code, sex, designation, basic pay, HRA, DA, deductions). The analyst determines the length and type of each field.
Key field (record key) One data item likely to be unique in all records of the file; used to identify the record for further processing.
File A collection of related records about the same entity (an employee file holds only employee records).
Database The highest level: a set of inter-related files for real-time processing; holds the data needed for problem-solving and can serve several users accessing data concurrently.

3.2 Types of files (pp.167–169). Master, transaction, table, report, back-up, archival, dump, library.

File type Course description
Master (3.2.1) The most important; essential records for maintaining the business; most file-design activity here. Reference master: records are static or change rarely (product file with descriptions and codes; customer file with name, address, account number). Dynamic master: frequently updated by transactions or order events. The two may be separate or combined (e.g. a sales ledger: reference data plus current transaction and balance).
Transaction (3.2.2) A temporary file used (1) to accumulate data about events as they occur; (2) to update master files. A transaction is any business event that affects the organisation and about which data are captured (purchases, hiring, sales).
Table (3.2.3) A special master file: permanent files of reference data used repeatedly in processing transactions, updating masters or producing output; data in tabular form; saves memory and eases program maintenance by keeping data out of programs/master records.
Report (3.2.4) Holds the contents of output reports or documents when a printer is not available (e.g. when input, processing and output run at the same time); written to tape or disk until printed. The process is spooling.
Backup (3.2.5) A copy of a master, transaction or table file made so a copy is available if anything happens to the original.
Archival (3.2.6) Copies made for long-term storage of data that may be needed much later; usually stored far from the computer centre.
Dump (3.2.7) A copy of computer-held data at a particular point in time: of a master file for recovery after corruption, or part of a program being debugged.
Library (3.2.8) Contains application programs, utility programs and system software packages.

3.3 File organisation (p.169). A file is organised so records are available for processing; look at both what data are recorded and how they are placed on the file.

3.3.1 Sequential (pp.169–171).

  • The simplest method; records stored and sorted physically one after another (tape or disk), without concern for the value of the data; records are stored by position, there are no addresses. Records can be added only at the end; you cannot insert in the middle without re-writing the file.
  • To read, the system starts at the beginning and reads one record at a time; it cannot jump to the fifteenth record directly. Records are in ascending or descending order by a key field, which occupies the same place in each record; the value being looked for is the search key (read the first record, compare, repeat until a match or end of file).
  • In a sequential update, transaction records are in the same sequence as the master, and records from both files are matched one at a time, giving an updated master.
  • Usually on magnetic tape, disk or mass storage.
Advantages (p.171) Disadvantages
Simple to understand Entire file must be processed even when the activity rate is low
Locating a record requires only the key record Transactions must be stored and placed in sequence before processing
Efficient and economical if the activity rate is high Timeliness of data deteriorates while batches accumulate
Relatively inexpensive I/O media and devices may be used Data redundancy is typically high (same data in several files sequenced on different keys)
Files are relatively easy to reconstruct (built-in backup)

3.3.2 Random or direct (pp.171–172).

  • Each record has a key field, but records need not be in key order; any record can be accessed independently of the others. The problem is turning the key into a storage location.
  • Direct addressing: the key field is the location; simple, but its requirements often prevent its use; rarely used.
  • Otherwise a formula or algorithm converts the key to a location; this is hashing (deriving a storage address from the record key). Algorithms named: division by prime, folding, extraction, squaring.
  • Division by prime (p.172): divide the key by a prime number and use the remainder (quotient discarded: modular division) as the storage address. If the key has many digits (e.g. 10), strip off the first or last 4 digits first. Course example: key 2345632278 → strip first 4 digits → 632278 → divide by 41 → quotient 15421, remainder 17 → address 17.
Advantages Disadvantages
Immediate access for inquiry and updating Online-file records are exposed to loss of accuracy; special backup and reconstruction procedures are needed
Immediate updating of several files from a single transaction Less efficient use of storage space than sequential
No sorting of transactions needed Adding and deleting records is more difficult than in sequential files
Relatively expensive hardware and software

3.3.3 Indexed (pp.173–174).

  • An index contains a record key and the storage address of the record. Searching an index is faster than scanning the data file.
  • It gives "the simplicity of a sequential file" plus direct access. Records are first stored sequentially by key; as they are recorded, the system builds one or more indexes linking key values with storage locations. To find a record: search the index (sequential scan of the index) for the key, take the address, access the record directly. Faster than sequential search, slower than direct addressing. Usually on magnetic disk or mass storage.
Advantages Disadvantages
Efficient and economical use of sequential processing when the activity rate is high Less efficient use of storage space than some alternatives
Quick access to records efficiently when activity is a small fraction of the total workload Access may be slower via indexes than with transform algorithms
Expensive hardware and software

Direct vs indexed (p.173). Direct files can be accessed only randomly; indexed files sequentially or randomly. Direct files use an algorithm to find a record; indexed files use an index.

Comparison of the three organisations (summary of pp.169–174):

Sequential Direct (random) Indexed (indexed sequential)
Order of records on file By key (ascending/descending) Any order (location from key by formula) Sequential by key, plus an index
Finding a record Read from the start, compare keys Compute the address from the key (hashing) Search the index, then go direct
Access Sequential only Random only Sequential or random
Best when Activity ratio high (batch) Immediate access, low activity, online Mixed: high activity runs and quick lookups
Medium Tape, disk or mass storage Disk Disk or mass storage
Weakness Whole file processed; slow to find; insert needs re-write Costly; loss-of-accuracy risk; harder to add/delete Storage use; slower than direct; costly

3.4 File design (pp.174–178). Basic factors in selecting the file medium and organisation method (six):

Factor Course content
Method of processing for updating files Batch: sequential processing on magnetic tape; the master file is re-created at every update; updates are done after a fairly large batch of transactions is accumulated (to cut set-up time). Random (direct): updates files as transactions occur; widely used for online processing through remote terminals. Online can also be used in batch mode to update several files simultaneously (e.g. a new open order file created from a batch of current orders while the customer and product files are accessed and updated).
Size of file Tape: convenient, economical for large files processed sequentially in batch. Disk: also holds large volumes, relatively expensive, suited to online use. Allow for growth potential.
File inquiry capabilities The ease of referring to a specific record without altering it. DASD (direct access storage devices) handle inquiries quickly; a teletype or CRT enters the inquiry and the answer returns within seconds; inquiries do not interrupt normal processing.
Response time The interval between entering an inquiry and getting the reply. Reservation systems (airline, hotel), stock quotations need fast response; DASD is most suitable.
Activity ratio The proportion of records processed in an updating run. More than 30 % = high → sequential processing is more economical. Less than 30 % = low → random/direct access is more suitable.
File volatility The additions, deletions and changes to a file; a file accessed often in a working period (banks, stock exchanges, airlines) is highly volatile.

Before designing a file the analyst records its aspects on a file work sheet, accompanied by detailed record layouts. Course work sheet items: file name and analyst; file update (batch/direct); file organisation (sequential/indexed sequential/direct); processing frequency (on demand, daily, weekly, monthly, other); activity ratio (low, moderate, high); direct access (yes, no, occasional); volatility (low, moderate, high); record characteristics (fixed or variable type, blocking factor); file dynamics (number of records, yearly additions, yearly deletions, percent growth); file media (tape, disk); sources of data; type of information required and reported.

3.5 Database design (pp.178–181). Historically files were built per application (payroll, personnel, accounts receivable, inventory), so the same data (e.g. employee name and address) appear in several files: redundancy, and a change of address must be processed several times, so data may become inaccurate. This is one reason database technology was developed.

  • Database: "a set of logically related files organised to facilitate access by one or more application programs and to minimise data redundancy"; a stored collection of data organised on the basis of relationships in the data rather than the convenience of storage structures. It is not a replacement for files.
  • Objectives (nine): eliminate redundant data as far as possible; integrate existing data files; share data among all users; incorporate changes easily and quickly; simplify the use of data files; lower the cost of storing and retrieving data; improve accuracy and consistency; provide data security against unauthorised use; exercise central control over standards.
  • DBMS (Database Management System): the set of programs for adding, modifying and retrieving data in a database. A database system merges data into one pool shared by all systems so a change automatically affects all relevant systems. Figure 4 (traditional systems): each system has its own data; Figure 5 (DBMS): data centralised.

Advantages of a database (p.180–181), seven:

# Advantage Course content
1 File consolidation Pooling reduces redundancy and inconsistency; a change in one system cascades to all others.
2 Program and file independence The definition of the files is separate from programs; programmers concentrate on program logic.
3 Cross/"Across" versatility Data can be retrieved in many ways: sequential access for reports in a prescribed order and random access for quick retrieval of one record.
4 Data security Password system limits access (read-only, write-only, specified records or fields).
5 Program development Programmers use standard means for data items.
6 Program maintenance Changes and repairs are relatively easy.
7 Special information Special-purpose report generators produce reports with minimum effort.

3.5.1 Logical and physical view of data (p.181). The logical view is what the data look like regardless of how they are stored; the physical view is how data exist in physical storage (how they are stored, accessed, related). Four views: three logical (the users' view, the programmer's view, the overall logical view = schema) and one physical.

3.5.2 Schema (pp.181–182). A coded set of tables, stored on direct access devices, that identifies each data item type, data aggregate, record and set by name or code, which item types make up aggregates and records, and which record types are members and owners of set types; it describes the data structure separately from the data itself. It is the overall logical data structure held by the DBMS; each time a program needs data the DBMS looks up the schema.

3.5.3 Sub-schema (p.182). Programmers are not allowed to write their own data division, for reasons of security or control; they are given a standard description of the logical data for the application, the subschema, which is similar to the schema in structure. The DBMS matches data requests made on a subschema with requests based on the schema.

3.6 Types of database (pp.182–184). In file terms, bytes → field → record → file; in database terms, bytes → data item or segment → data entry (collection of segments) → data set (series of data entries) → the complete collection of data sets is the database. Databases relate data sets in one of three models:

Model Structure Course points
Hierarchical (3.6.1, Fig. 6) Tree: data get more detailed further out on the branches; each segment (node) can have two or more subordinate nodes. Each node has only one parent.
Network (3.6.2, Fig. 7) Like hierarchical, but... A node may have more than one parent; a hierarchical DBMS is a subset of a network DBMS; trade-off between the simplicity of hierarchical design and the storage efficiency of network structure.
Relational (3.6.3, Fig. 8) Two-dimensional tables of rows and columns; no parent or node data sets. A table is a relation; a row (record) is a tuple; a column (field) is an attribute name or domain.

Relational DBMS basic operations (7): create or delete table; update, insert or delete rows; add or delete columns; copy data from one table into another; retrieve or query a table, row or column; print, recognise or read a table, row or column; join or combine tables on a value. Strengths: great flexibility, high data security, little memory or secondary storage. Weaknesses: creating tables is elaborate; more time to access information than the other two (much more information to search); some implementations use a fixed amount of storage per field (insufficient storage utilisation). Despite these, it is the most popular structure. Example: an inventory of 1600 tuples with 5 attributes = a table of 1600 rows and 5 columns.

3.7 Coding system (p.185). Information systems are designed with space, time and cost savings in mind; coding systems reduce input, control errors and speed up the process. A code is "an ordered collection of symbols designed to provide unique identification of an entity or attribute" (a brief number, title or symbol); its main purpose is to facilitate identification and retrieval of items of information.

3.8 Types of codes (pp.185–187). Seven:

Code Meaning Course example
Classification code (3.8.1) Puts entities (events, people, objects) into distinct classes; a code identifies one class from another; a single-digit code simplifies input, removing lengthy descriptions and judgements.
Function code (3.8.2) States the activities or work to be done without narrative; frequent in transaction data to tell the system how to process the data. A = add, D = delete, U = update, E = edit; or 1 = addition, 2 = modification, 3 = deletion
Card code (3.8.3) Lets the program distinguish types of card and check that a card's contents are correct; used for punched cards only (batch).
Sequence code (3.8.4) Numbers or letters assigned in series; show the order in which events occurred; simple; can cover an unlimited number of items with few digits; new items get the next higher unused number; no provision for classifying like items. Employee numbers issued consecutively as hired
Significant-digit subset code (3.8.5) Sub-codes within a larger number; each part has a meaning. Prefix PL = plastic, ST = steel; examination number 11 02 1978: 11 = centre, 02 = post number, 1978 = number assigned to the candidate
Mnemonic code (3.8.6) Letters, numbers, or both, derived from the description of the item; fewer errors than other codes where items are few and stable. M and F more reliable than 1 and 2; Y/N; TV-CL-16 (16-inch colour TV); cm, m, km
Acronym (3.8.7) A mnemonic formed from the first letter or letters of several words; may become a word itself. MIT (Master of Information Technology); RADAR (Radio Detecting and Ranging)

Code and worked examples

Example 1 — Division by prime (Course p.172; checked). Key 2345632278.

  1. The key has 10 digits, so strip the first 4: 2345 | 632278.
  2. Divide by the prime 41: 632278 ÷ 41 gives quotient 15421, since 41 × 15421 = 632,261; remainder 632,278 − 632,261 = 17.
  3. Discard the quotient; the remainder is the address: record goes to location 17.

Variants (not in the course): stripping the last 4 digits instead (allowed by "first or last") gives 234563; 234563 = 41 × 5721 + 2, so the address would be 2. Two different keys can give the same address (632278 and 632319 = 632278 + 41 both give 17); the course does not discuss such clashes.

Example 2 — Sequential search (Course p.170). The file holds employee numbers 1200 to 1250 in order (51 records); the search key is 1234.

Step Record read Compare with 1234 Result
1 1200 1200 ≠ 1234 read next
2 1201 ≠ read next
... ... ... ...
35 1234 = found

The record is the 35th (1234 − 1200 + 1), so 35 reads. If the key were 1260 (not in the file), all 51 records would be read and the search would end at the end of file. A direct or indexed file would not read from the start.

Example 3 — Sequential batch update (Course pp.166, 174–175; keys are our illustration). Old master (father) in key order: 101, 105, 110, 120. Transactions (sorted in the same order): 105 modify, 110 delete, 115 add, 130 add.

Step Master record read Transaction read Action New master (son) so far
1 101 105 master key < transaction key: copy 101 101
2 105 105 keys match, modify: write updated 105 101, 105*
3 110 110 match, delete: do not write 101, 105*
4 120 115 transaction key lower: add 115 101, 105*, 115
5 120 130 master key lower: copy 120 101, 105*, 115, 120
6 end 130 add 130 101, 105*, 115, 120, 130

Course points: the file being read is the brought-forward (old master, "father") file; the new one is the carried-forward ("son") file; the previous one ("grandfather") is kept for recovery; the transaction file is in the same sequence as the master; the whole file is processed even if only a few records change (high cost at low activity).

Example 4 — Activity ratio (Course p.176). Rule: more than 30 % of records updated in a run = high → sequential; below 30 % = low → direct.

File Records Updated per run Activity ratio Better method
Monthly payroll 10,000 4,200 42 % (> 30 %) Sequential (batch)
Airline seats 10,000 800 8 % (< 30 %) Direct (random), DASD

Example 5 — Record layout and length (Course p.167).

Data item Type Length Decimal
Name C 20
CODE N 5
SEX C 1
DESIGNATION C 15
BASIC N 8 2
HRA N 6 2
DA N 6 2
DEDUCTION N 6 2

Eight data items. Total length = 20 + 5 + 1 + 15 + 8 + 6 + 6 + 6 = 67 characters (assuming, as in this layout style, that each length already includes the decimal places). C = character, N = numeric. The key field would be CODE (the item likely to be unique; the course gives the principle, not this choice).

Example 6 — Relational query (Course Fig. 8, p.184). "Who supplies the blue chairs?"

Step Table Look at Found
1 FURNITURE Type = Chair and Colour = Blue Product 2892, supplier number 30
2 SUPPLIER Supplier number 30 BALOGUN (purchases this year 26,035.00)

This is a join on the supplier number. FURNITURE has 5 tuples and 5 attributes (product number, type, colour, quantity in stock, supplier number); SUPPLIER has 3 tuples and 3 attributes. Note for exam traps: FURNITURE also lists suppliers 133 and 150, which are not rows in the printed SUPPLIER table (only 30, 26, 159 appear).

Common mistakes

  • Master vs transaction vs backup vs archival.
File Key word
Master permanent business records (reference or dynamic)
Transaction temporary; records events; used to update the master
Backup copy so a copy exists if the original is damaged
Archival copy for long-term storage, kept far from the computer centre

Also: a dump is a copy at a point in time (recovery or debugging); a table file is permanent reference data; a library file holds programs.

  • Sequential vs direct vs indexed. Sequential: read from the start, add at the end. Direct: algorithm/hashing to compute the address; random access only. Indexed: records in sequence plus an index (key and address); access sequentially or randomly; slower than direct addressing.
  • Hashing vs index. Hashing computes an address from the key; an index looks it up in a table of key/address pairs.
  • High vs low activity ratio. More than 30 % = high = sequential; less than 30 % = low = direct.
  • Brought-forward vs carried-forward. The file being read is brought-forward (old master); the file being written is carried-forward (new master). Grandfather-father-son = three generations.
  • Schema vs sub-schema. Schema = overall logical structure held by the DBMS; sub-schema = the standard logical description given to a programmer for one application. Logical views are three (users', programmer's, schema); physical is one.
  • Hierarchical vs network vs relational. Hierarchical = one parent per node (tree); network = more than one parent; relational = tables (relation, tuple, attribute). A tuple is a row (record), an attribute a column (field).
  • Database vs file. A database is not a replacement for files: it is a set of logically related files organised to cut redundancy.
  • Mnemonic code vs acronym vs sequence code. Mnemonic: letters/numbers derived from the description (TV-CL-16, M/F). Acronym: initials of several words (RADAR). Sequence: just the next number in series, no classification.

Errata

Page Course says Correct Note
p.164 Unit heading "UNIT 4 INPUT DESIGN AND CONTROL" File and Database Design (module contents p.112: "Unit 4 File and Database Design"; the unit content is files, databases, codes) Heading misprinted; copy of Unit 2's title. Use "File and Database Design".
pp.174–175 "Fig. 2 illustrates this method" (random processing), but the figure below is captioned "Fig. 3: Updating Direct Access File"; the next figure is also "Fig. 3: Simultaneous Updating of Multiple Files". Direct-access update is Fig. 2; simultaneous updating is Fig. 3 Figure caption numbered 3 twice.
p.165 Intro: file organisation "may be sequential, index sequential, inverted list or random" Section 3.3 and the self-assessment (p.178, "three methods") cover three: sequential, random or direct, indexed "Inverted list" is named once and never explained.
p.182 "The DMBS has the job of matching data requests…" DBMS Typo.

Practice questions

  1. (Self-assessment i) Fill in the file hierarchy: data item → ______ → ______ → database.
  2. (Self-assessment ii) Which of these is not one of the types of file listed in the course? A. Master B. Transaction C. Index D. Archival

  3. (Self-assessment iii) The three methods of file organisation are ______, ______ (random) and ______.

  4. (Self-assessment iv) The basic factors in selecting file media and organisation are: method of processing for updating, size of file, file inquiry capabilities, activity ______ of records, file ______ and the response time.
  5. Output that cannot be printed when produced is written to a file on tape or disk until a printer is free. This process is called: A. hashing B. spooling C. dumping D. archiving

  6. Key 632278 is divided by 41, giving quotient 15421. The storage address under the division-by-prime method is: A. 15421 B. 41 C. 17 D. 632278

  7. A file has an activity ratio of 45 %. According to the course, the more suitable method is: A. direct access B. sequential processing C. hashing D. spooling

  8. In the hierarchical model each node has: A. any number of parents B. exactly one parent C. no parent D. one child

  9. In a relational database a row is called a ______ and a column an ______.

  10. (TMA 4) The registration number 11021978 is built from the centre (11), the post (02) and the candidate number (1978). This is a: A. sequence code B. acronym C. significant-digit subset code D. function code
Answersshow
  1. Record; file. 3.1.1.
  2. C. The eight types: master, transaction, table, report, back-up, archival, dump, library.
  3. Sequential; direct; indexed. 3.3.1–3.3.3.
  4. Ratio(s); volatility. 3.4 (six factors).
  5. B. 3.2.4: spooling into a report file.
  6. C. 632278 − 41 × 15421 = 17; the remainder is the address.
  7. B. More than 30 % = high activity ratio → sequential (p.176).
  8. B. 3.6.1: each node has only one parent.
  9. Tuple; attribute (attribute name or domain). 3.6.3.
  10. C. 3.8.5 significant-digit subset code.

External resources

Each URL below was opened and returned HTTP 200 on 2026-11-22.

  • Wikipedia, Hash function (the division method used in Unit 4): https://en.wikipedia.org/wiki/Hash_function
  • Wikipedia, Relational model (relation, tuple, attribute): https://en.wikipedia.org/wiki/Relational_model
  • Wikipedia, Database index (what an index is, for indexed files): https://en.wikipedia.org/wiki/Database_index
  • Wikipedia, Backup rotation scheme (grandfather-father-son generations): https://en.wikipedia.org/wiki/Backup_rotation_scheme
  • Wikipedia, Data dictionary (Unit 2, section 3.6): https://en.wikipedia.org/wiki/Data_dictionary

Module 3 of 3 · 4 units · PDF pp.189–257

Module 3 — Development, Quality and Implementation

Page convention. "p.N" is the PDF page N of CIT212.pdf. The printed page number in the page header is the PDF page minus 18 (for example PDF p.189 is printed 171). Module 3 covers PDF pp.189–257: Unit 1 pp.189–206, Unit 2 pp.207–221, Unit 3 pp.222–234, Unit 4 pp.235–257. Figure and table numbers (Fig. 1–7, Tables 1–2) are the course's own; Units 2 to 4 restart their figure numbering.

Likely to be examined. Each unit has an "Objectives" list (Unit 1 p.190, Unit 2 p.208, Unit 3 p.223, Unit 4 p.237). Self-assessment exercises: Unit 1 p.195 (3 questions); Unit 2 p.212 (Exercise 1, 3 questions) and p.218 (Exercise 2, 3 questions); Unit 3 p.227 (Exercise 1, 3 questions) and pp.230–231 (Exercise 2, 3 questions); Unit 4 p.242 (3 questions). Tutor-marked assignments: Unit 1 p.206 (2), Unit 2 p.221 (3), Unit 3 p.234 (2), Unit 4 p.257 (3). All of these are answered in the Practice questions of the matching unit. The exam is computer-based (MCQ and fill-in-the-blank), so know the named lists exactly (quality factors, test types, documentation types, conversion methods, contract clauses, vendor criteria) and the look-alike pairs (verification vs validation, unit vs system testing, parallel vs pilot vs phase-in, renting vs leasing vs purchase).

Jump to unit1234

Module 3 · Unit 1 of 4

Unit 1 — System Development

What you must know

  • List the nine major tasks of system development (implementation planning to user acceptance review) and the three parts of an implementation plan.
  • Explain what a prototype is, its five characteristics and its two uses.
  • State the six factors to consider before selecting a system, the five selection steps, and the ten criteria for selecting software.
  • Define "benchmark" and say what benchmarking validates and what limits it.
  • Name the three control structures (sequence, if-then, do-while) and read a program flowchart or pseudocode built from them.
  • Choose a programming language (FORTRAN, COBOL, BASIC, C, dBASE) from the five selection parameters.

Key ideas

Intuition. Design says what the system must do; development builds it. This unit covers what has to happen in the build phase (plans, code, tests, manuals, training), how to buy hardware and software wisely, how to compare machines (benchmark), and how to plan program logic before coding.

The development phase (pp.189–190). "The fourth phase in the life cycle of a system is the software development." The system is developed "to conform to the design specification prepared in the preceding phase". It involves heavy expenditure (extra staff, machines, materials, computer facilities). The principal activities are (a) external system development and (b) internal system implementation planning, preparation of manuals and personnel in-house training, and equipment acquisition and installation. "Software development and performance testing are considered to be the principal activities of internal system development." (The printed sentence is garbled; see Errata.)

3.1 Tasks of system development (pp.190–192). The nine tasks, in the course's order:

# Task What the course says
(i) Implementation planning Starts once the development phase is approved. Essential parts: (a) a plan for testing the programs, both as an integrated assembly and as an element of the whole business system; (b) a plan for training the people who provide inputs, receive outputs, operate or maintain the system; (c) a conversion plan for procedures, programs and files before the changeover from old to new.
(ii) Software development phase Runs alongside implementation planning. System flowcharts are expanded with more detail for program components, the complete database is developed, input and output files are identified, and program logic flowcharts are prepared for each program component.
(iii) User review Reviews with the principal user throughout. Test, training and conversion plans must be reviewed; user concurrence with the implementation plan is extremely important.
(iv) Equipment acquisition and installation Special hardware is identified in design; if not yet ordered, order now. Not all hardware is needed at once, so prepare an acquisition schedule.
(v) Coding, debugging and testing of computer programs Each program is coded and debugged: compiled error-free and run successfully on the programmer's test data.
(vi) System testing Verifies that the computer-based system has met its design objectives.
(vii) Reference manual preparation Manuals for everyone associated with the new system.
(viii) Personnel training Operating, programming and user staff trained with the manuals, forms and procedures as training aids; all training is completed before the user acceptance review.
(ix) User acceptance review At the end of the phase, management of the user organisation reviews the system, with representatives of the information service and other affected organisations. The design phase report, development phase report and test reports are the key documents.

3.2 Prototype installation (pp.192–193). "A prototype installation is the process of creating, developing and refining a working model of a final system." It does not contain all the features or functions of the final system, but enough that people can use it to decide what they like and dislike and which features to add or change. Application prototyping is the process of developing and using the prototype.

Characteristics (the course's list):

  • (i) a live, working application;
  • (ii) its main purpose is to test the assumptions made by analysts and users about the required features;
  • (iii) can be created quickly;
  • (iv) follows an iterative process;
  • (v) relatively cheap.

Two primary uses: (1) an effective device for clarifying user requirements (instead of relying only on written specifications); (2) verifying the feasibility of a system design (experimenting with different application characteristics and evaluating user reaction). Rationale: the need to design and develop information systems quickly, efficiently and effectively (pp.192–193). The summary (p.206) adds: prototyping suits cases where predicting the user's requirement is not possible.

3.3 Hardware and software selection (pp.193–195). Acquiring hardware or software is "handled in the same way as any other business decision" (p.193).

Factors to consider before system selection (p.193):

  • (a) define system capabilities that make sense for the business;
  • (b) specify the magnitude of the problem (a few peripherals or a mainframe);
  • (c) assess the competence of the in-house staff;
  • (d) hardware and software should be considered as a package;
  • (e) develop a time frame for the selection process;
  • (f) provide user indoctrination, especially for first-time users (selling the system to user staff, training them, creating an environment conducive to implementation).

The selection process is a project: form a project team with management's help. Five steps (p.194):

Step Name Content
1 Requirement analysis Understand the user's requirements within the organisation's objectives and the installation environment.
2 System specifications Clearly defined; reflect the actual applications; include system objectives, flowcharts, input-output requirements, file structure and cost.
3 Request for proposal Sent to selected vendors for bidding after steps 1 and 2.
4 Evaluation and validation Evaluation ranks the vendor proposals (price, availability, technical support) and finds the best fit. Validation checks that the vendor can match its claims, especially system performance.
5 Vendor selection Best combination of reputation, reliability, service record, training, delivery time, lease/finance terms. Selected vendors present their systems; the chosen one goes through contract negotiation before implementation.

Software selection (pp.194–195). Two ways to acquire software: custom-made or "off-the-shelf" package. Packages are in demand because they are quite cheap.

Advantages of a package Drawbacks of a package
1 A good package gets the system running quickly. May not meet user requirements in all respects.
2 MIS personnel are released for other projects. Extensive modification usually costs the vendor's support.
3 "Home-grown" software takes more time and its cost cannot be predicted.
4 A package can be tested before purchase.

"Price alone cannot determine the quality of software": the project team sets selection criteria (p.195). The ten criteria:

Criterion Meaning (course wording)
(a) Reliability consistent results without failure for a specified time period
(b) Functionality functions to standards
(c) Capacity satisfies the user's volume requirements
(d) Flexibility adapts to changing needs
(e) Usability user-friendly
(f) Security maintains integrity and prevents unauthorised use
(g) Performance delivers the results as required
(h) Serviceability good documentation and vendor support
(i) Ownership right to modify and share use of the package
(j) Minimal costs justified and affordable for the intended application

3.4 Benchmark testing (pp.196–197). The word comes from a machinist who measured parts at each bench. To compare systems, run the same set of "benchmark" programs through each. Definition (p.196): "A benchmark is a sample program specifically designed to evaluate the performance of different computers and their software." Needed because computers do not generally use the same instructions, words of memory or machine cycle for a given problem. For software, benchmarking mainly validates the vendor's claims about:

  • the minimum hardware configuration needed to operate a package;
  • the time to execute a program in an ideal environment, and how performance of your package and of other programs is affected when running in multiprogramming mode.

Limits: "The more elaborate the benchmarking, the more costly is the evaluation." Keep the user's goals in mind; time constraints limit thoroughness; compromise on how much to test while still ensuring the functional criteria are met. Benchmarks run in almost any environment (batch, online, users linked directly or by telecommunications). Common benchmarks test central processor speed with typical instructions and multiple job streams in multiprogramming. They can also be built on an expected language mix, a mix of programs with widely varying input and output volumes, and terminal response time. It is "one of the evaluation techniques used by the computer purchasers" to decide which machine best meets their speed and cost requirements (p.197).

3.5 Preparing the software development cycle (pp.197–202). Programming starts after systems design is completed; work done in design is the foundation, and results "may be disastrous" if structured design principles were not followed. The structure chart (Fig. 1, p.197, a Sales Order Entry system) shows a system decomposed into related modules, each a self-sufficient function (the figure's modules: Data Entry, Data Validation, Customer credit check, Shipping documents, Sales Report, Add New Accounts, Modify Existing Data). The modular approach plus programming guidelines plus regular reviews are the fundamental concepts of structured programming.

  • 3.5.1 Identifying programs (p.197). Every large system consists of a number of small programs, easier to write, modify, troubleshoot and maintain than one very large one. The modular approach is the basis for identifying separate programs.
  • 3.5.2 Program logic and flowcharts (p.198). Before coding, the detailed logic of each program is shown as program flowcharts, data flow diagrams or a more English-like form called pseudocode. A program flowchart is "a detailed graphical representation of the logical flow of data within that module", the "logical road map" for the programmer, built from standard symbols (Fig. 2): processing, input/output, decision-making, terminal, connector, off-page connector.
  • 3.5.3 Control structures (pp.198–201). Any program can be built from a few basic structures. The course says "three basic control structures" but then names two in the sentence; the three it describes are:
Structure Use Course example
Simple sequence (p.199) A series of steps in linear order, from first to last. Fig. 3: validate employee wage data (Enter wage data → Validate wage data → Print error report → END).
If-then (p.199) Transfers control on a condition. Fig. 4: employees with more than 10 years of experience are listed (decision "more than 10 yrs of exp?": Yes → print employee name; No → skip).
Do-while (p.200) Loops: repeats a sequence while a pre-specified condition holds. The test is made immediately on entry; processing inside the loop happens only after the test, on the line that returns to the top. Therefore a record must be read before the do-while is entered the first time. Fig. 5: read each employee record and print it.

Control structures combine into complete programs (p.201). Fig. 6 is a complete flowchart that reads accounts-receivable records and prints customers whose account is more than 60 days overdue and whose balance is greater than N1000.00 (the figure prints "Rs. 10,000"; see Errata).

  • 3.5.4 Pseudocode (p.201). Expresses program logic "in English statements", verbal rather than graphic. Advantages: (1) it closely resembles the form the actual code will take; (2) it avoids laying out symbols on paper. Fig. 7 (p.202) pairs a flowchart with this pseudocode: Read a record; Do-while more data; Write record; Read record; End do; END.

3.6 Software specification language selection criteria (pp.202–205). After flowcharts or pseudocode, a language is chosen. Often the choice follows the language of most of the organisation's other programs (consistency, easier maintenance). Languages discussed:

Language Meaning Strengths Weaknesses
FORTRAN FORmula TRANslation Compact; serves scientists and economists; large library of mathematical and engineering subroutines. Handling files with voluminous data is not a strength.
COBOL COmmon Business Oriented Language Popular for large data-processing problems; handles large files; programs easy to understand and edit; supports indexed files; no limit on files open. Simple arithmetic only; does not help with complex mathematical problems.
BASIC Beginners' All-purpose Symbolic Instruction Code Most popular conversational language; simple; suits mathematical and business problems; designed for time-sharing, usable in batch. Most versions do not support indexed files.
C (name) Powerful and compact; more control over hardware (features normally only in assembly); popular for system software (operating systems, utilities). Program not easily decipherable; takes a long time to learn.
dBASE Fourth Generation Language More an application development tool than a language; stores and retrieves data; quick query and reporting facility; simple. If the volume of data is very large, the program can become quite slow.

The five parameters for language selection (p.203): volume of data, complexity of processing, compatibility with other systems, types of input/output, development efforts.

Parameter What the course says
3.6.1 Volume of data (pp.203–204) Two aspects: number of files and number of records. Some languages limit files open at once (dBASE III+ can open ten data files; COBOL has no such restriction). Most languages slow down as file size grows, especially fourth-generation languages. Some tools (DBMS-based products) are not desirable for very low data volume.
3.6.2 Complexity of processing (p.204) Simple reporting programs need little computation. Heavy computation (course example: analysing data from a drilling rig exploring for oil) benefits from mathematical co-processors; FORTRAN is desirable, some spreadsheets also help. Repetitive but not complex computation: COBOL or dBASE.
3.6.3 Compatibility with other systems (p.204) A new system must co-exist with others; some data files may be accessed by both old and new systems. Different languages may be unable to read the same files (COBOL programs cannot directly read a dBASE III file); two programs may need to invoke each other.
3.6.4 Types of input/output (pp.204–205) Two issues: types of I/O device (usually yes/no: for plotter output dBASE may not be suitable) and complexity of formats (graphic output: integrated packages like LOTUS 1-2-3 or VP Planner; complex screens or printouts: ask how critical the format is and whether it can be modified to simplify programming; example: dBASE REPORT FORM prints the page number at the top, a user may want it at the bottom; ad hoc queries suit a powerful query language like dBASE III).
3.6.5 Development efforts (p.205) Can be the decider. Different effort for the same task because of (i) features available (most BASIC versions lack indexed files, COBOL has them) and (ii) learning time (a language the team does not know). A language is often chosen because people in the organisation are familiar with it.

Summary points (p.206). The system analyst develops information systems useful to management and employees. Prototyping is appropriate when predicting the user's requirement is not possible; a prototype has the essential features but not necessarily all details of the user interface or performance efficiency.

Code and worked examples

There is no programming language in this course; the "code" is flowchart logic and pseudocode. Traces below were checked by hand (and the Fig. 6 filter was re-checked with a script).

Example 1 — Simple sequence and if-then (Course example p.199, Figs. 3–4). Fig. 3 sequence (steps run once, in order):

Text
START Module 1
  1. Enter wage data
  2. Validate wage data
  3. Print error report
END

Fig. 4 if-then on the employee list (more than 10 years):

Employee Years Years > 10? Action
Ada 12 Yes print name
Bayo 10 No (10 is not more than 10) skip
Chidi 4 No skip
Dayo 11 Yes print name

Printed: Ada, Dayo. The condition is strictly "more than".

Example 2 — Do-while and pseudocode (Course example p.200 Fig. 5; p.202 Fig. 7). Pseudocode (Fig. 7):

Text
Read a record
Do-while more data
    Write record
    Read record
End do
END

Trace on a file of three records R1, R2, R3 then end-of-file:

Step Action "More data?" Output
1 Read R1 (before the loop)
2 Test on entry Yes (R1 held)
3 Write R1, read R2 R1
4 Test Yes
5 Write R2, read R3 R1, R2
6 Test Yes
7 Write R3, read next = end of file R1, R2, R3
8 Test No exit loop, go to END

Without the first "Read a record" before the loop, the first test would find no record, which is why the course insists a record be read before the do-while is entered.

Example 3 — Complete program flowchart (Course example p.201 Fig. 6). Rule: print the customer name when days overdue > 60 AND balance > N1000 (the course text uses N1000.00; the figure prints Rs. 10,000, see Errata). Using N1000 as in the text:

Customer Days overdue Balance (N) > 60 days? > 1000? Printed?
A 75 1,500 Yes Yes Yes
B 90 800 Yes No No
C 45 5,000 No (not tested) No
D 61 1,000 Yes No (equal, not greater) No
E 120 2,300 Yes Yes Yes
F 60 9,000 No (equal, not more) (not tested) No

Output list: A, E. In the flowchart the two decisions are in series, so a "No" on either one goes to "Read next record", and the loop ends at "End of file?".

Example 4 — Choosing a language from the five parameters (course criteria pp.202–205).

Situation Best choice Course reason
Analyse seismic data from an oil rig, heavy computation FORTRAN complexity of processing; mathematical/engineering subroutine library
Payroll with very large files and repetitive arithmetic COBOL handles large files; repetitive computation; supports indexed files
Operating-system utility needing hardware control C system software; features like assembly language
Quick report from a small data set, many ad hoc queries dBASE query and reporting facility, 4GL
Program needs indexed files not BASIC (most versions lack them) development effort / feature availability
Old programs are all COBOL, new module must read their files COBOL (or arrange program-to-program calls) compatibility; COBOL cannot directly read a dBASE III file

Example 5 — Benchmark comparison (illustrative numbers; the course gives none). Same three benchmark programs run on two computers (seconds):

Program Computer X Computer Y
Sales order batch 12 10
Payroll update 30 36
Report print 8 7
Total 50 53

Reading: Y is faster on two programs, but X is faster on the heavy payroll job, so X wins on total time. Because the same programs ran on both, the differences come from the machines (the course's point on p.196). Benchmarking also checks vendor claims such as the minimum configuration, and a more elaborate benchmark costs more (p.196).

Common mistakes

  • Benchmark vs prototype. A benchmark is a sample program that measures performance of hardware or software; a prototype is a working model of the system used to clarify requirements. Do not swap them.
  • Evaluation vs validation (step 4 of selection). Evaluation ranks proposals (price, availability, technical support); validation checks the vendor's claims, especially system performance. Both belong to the same step.
  • Control structures. The three are simple sequence, if-then, do-while. An MCQ may offer "case" or "for" as the third; the course names sequence. A do-while tests first, so read a record before entering it.
  • Language traps. FORTRAN = scientific, weak on voluminous files; COBOL = large business files, weak on complex mathematics; BASIC lacks indexed files; C = system software and hard to learn; dBASE = 4GL, query and reports, slows with large data. COBOL cannot directly read a dBASE III file.
  • Pseudocode is verbal; a flowchart is graphic. Do not call pseudocode a diagram.
  • Selection steps vs selection factors vs software criteria. Factors (six, p.193), steps (five, p.194), criteria (ten, p.195) are three different lists.

Errata

Page Course says Correct Note
198 "can be developed using only three basic control structures; (i) if-then and (ii) do-while" Three structures: simple sequence, if-then, do-while Only two are named in the sentence; the third (simple sequence) is described on p.199 (Fig. 3).
201 Text: balance "greater than N1000.00"; Fig. 6 decision box: "Is Balance > Rs. 10,000" One threshold The figure disagrees with the text and uses a different currency and amount. For the exam use the text: N1000.
191 "compiled error free ... using the rest data" (printed "complied ... rest data") compiled ... test data Typographical errors in task (v).
189–190 "(a) External system development and (b) Internal system implementation planning, preparation of manuals ..." Garbled sentence across a page break Meaning from the next sentence: software development and performance testing are the principal activities of internal system development.

Practice questions

  1. (MCQ) The course calls which phase of the system life cycle "the fourth"? A. Feasibility study B. System design C. Software development D. Post-implementation review
  2. (MCQ) Which is not one of the course's characteristics of application prototyping? A. It is a live working application. B. It follows an iterative process. C. It is relatively cheap. D. It contains every feature of the final system.
  3. (MCQ) In the five steps of hardware selection, ranking vendor proposals and checking that the vendor can match its claims is the step called: A. Request for proposal B. System specifications C. Evaluation and validation D. Vendor selection
  4. (MCQ) A sample program designed to evaluate the performance of different computers and their software is a: A. prototype B. benchmark C. pseudocode D. run book
  5. (Fill) A do-while structure tests its condition ______ on entry, so a ______ must be read before the structure is entered for the first time.
  6. (MCQ) Which language does the course say lacks indexed files in most versions? A. COBOL B. FORTRAN C. BASIC D. dBASE
  7. (Fill, TMA 2) List the five parameters for choosing a programming language: volume of data, ______, compatibility with other systems, ______, ______.
  8. (Self-assessment i) List the major tasks of system development.
  9. (Self-assessment ii) What factors must be considered before system selection?
  10. (Self-assessment iii / TMA 1) Explain briefly the criteria for software selection. Describe benchmark testing briefly.
Answersshow
  1. C: "The fourth phase in the life cycle of a system is the software development" (p.189).
  2. D: a prototype "does not contain all the features" of the final system (p.192).
  3. C: evaluation ranks proposals, validation checks vendor claims (p.194).
  4. B: definition on p.196.
  5. Immediately (first); record. A do-while allows no processing between entry and the test (p.200).
  6. C: "most versions of this language do not support indexed files" (p.203; again p.205).
  7. complexity of processing; types of input/output; development efforts (p.203).
  8. Implementation planning; software development phase; user review; equipment acquisition and installation; coding, debugging and testing of programs; system testing; reference manual preparation; personnel training; user acceptance review (pp.190–192).
  9. Define capabilities that make sense for the business; specify the magnitude of the problem; assess in-house staff competence; treat hardware and software as a package; develop a time frame; provide user indoctrination (p.193).
  10. Criteria (p.195): reliability, functionality, capacity, flexibility, usability, security, performance, serviceability, ownership, minimal costs. Benchmark testing (pp.196–197): run the same sample programs through each system to compare performance; for software it validates the vendor's claims on minimum hardware configuration and on execution time and multiprogramming effects; more elaborate benchmarking costs more, and time limits how thorough it can be.

Module 3 · Unit 2 of 4

Unit 2 — System Control and Quality Assurance

What you must know

  • Define quality assurance and state its goal at each stage of the software life cycle.
  • List the twelve quality factors and define the common ones (correctness, reliability, maintainability, portability, and others).
  • Name the three levels of quality assurance (testing; verification and validation; certification) and say what each does.
  • Distinguish error avoidance, error detection and correction, and error tolerance; list the three types of maintenance and their relative frequency.
  • Describe unit testing, the five steps of system testing, and the six special system tests; distinguish live from artificial test data.
  • State the objectives of system control, internal vs external control, and define an audit trail.

Key ideas

Intuition. Software is too big to be right by luck. Quality assurance (QA) is the set of checks that run through the whole life cycle; testing looks for failures; controls keep the running system accurate, secure and traceable.

QA defined (pp.207–208). "Basically quality assurance is the review of software products and its related documentation for completeness, correctness, reliability and maintainability." Software development strategies lack a standard, so software often misses its application objectives; hence controls must ensure a quality product.

3.1 QA in the software life cycle (pp.208–209). Each stage has a QA goal.

3.1.1 Quality factors specifications (pp.208–209) — the twelve factors:

# Factor Course definition
(i) Correctness extent to which a program meets system specifications and user objectives
(ii) Reliability degree to which the system performs its intended functions over time
(iii) Efficiency computer resources required by a program to perform a particular function
(iv) Usability effort required to understand and operate the system
(v) Maintainability ease with which program errors are detected and removed
(vi) Testability effort required to test a program to ensure its correct performance
(vii) Portability ease of transporting a program from one hardware configuration to another
(viii) Accuracy required precision in input, editing, computations and output
(ix) Error tolerance error detection and correction versus error avoidance
(x) Expandability ease of expanding the existing database
(xi) Access control and audit control of access to the system and the extent to which that access can be audited
(xii) Communicativeness usefulness and effectiveness of the inputs and outputs
Stage QA goal (p.209)
3.1.2 Software requirements specifications Generate the requirements documents that give the technical specifications for developing the software.
3.1.3 Software design specifications The design document defines the overall architecture providing the functions and features in the requirements document.
3.1.4 Software testing and implementation Testing: ensure completeness and accuracy of the system and minimise retesting. Implementation: provide a logical order for creating the modules and so the system.
3.1.5 Maintenance and support Software work needed for the system to keep complying with the original specifications; goal: a procedure for correcting errors and enhancing software.

3.2 Levels of QA (p.210). "Analysts use three levels of quality assurance: testing, verification with validation and certification."

Level What it is
3.2.1 Testing Expensive and time-consuming. Users think it proves a program error-free, but analysts cannot prove that. The practical view: "testing is the processing of executing a program with the explicit intention of finding errors, that is, making the program fail. A successful test, then, is one that finds an error."
3.2.2 Verification and validation Verification: like testing, intended to find errors; done by executing a program in a simulated environment. Validation: using software in a live environment to find errors. Commercial software meant for dealers first goes through verification, "sometimes called alpha testing"; feedback leads to changes; then selected users run it live at beta test sites, processing live transactions and producing normal output. Validation may continue for several months and failures lead to further changes.
3.2.3 Certification The last level: certify that the package conforms to standards. A team of computer specialists tests and reviews how well it meets user requirements and the vendor's claims; issued only if all tests are passed. "Certification, however, does not mean that is the best package to adopt. It only attests that it will perform what the vendor claims."

3.3 Design objectives: reliability and maintenance (pp.211–212). The two operational design objectives are reliability and maintainability.

Reliable system (p.211): "does not produce dangerous or costly failures during its normal use." Two levels: (1) the system meets the right requirements (needs thorough requirements study); (2) the actual working of the delivered system (tied to software engineering and development). Table 1, three approaches to reliability:

# Approach Description Example (course)
1 Error avoidance Prevents errors from occurring in the software "It is impossible in large systems."
2 Error detection and correction Recognises errors when they are encountered and corrects the error or its effect so the system does not fail Traps and modifies illegal arithmetic steps; compensates for unexpected data values
3 Error tolerance Recognises errors when they occur but lets the system keep running with degraded performance Shuts down part of the system; does not perform some processing but keeps the system operational

Maintainable systems (p.212): average life of a system is generally 4 to 6 years, with the oldest application often over 10 years; so analysts control the need for maintenance through design and testing and provide the ability to maintain through proper design practice.

3.4 Maintenance issues (pp.212–213). Studies reveal: (a) 60 to 90 percent of the overall cost of software during the life of a system is spent on maintenance; (b) maintenance often is not done efficiently; (c) software demand grows faster than supply, and programmers spend more time on maintenance than on new development. Table 2 (p.213), types of system maintenance (from Lientz and Swanson, 1980):

Category Activity Relative frequency
Corrective Emergency fixes, routine debugging 20%
Adaptive Accommodation of changes to data and files and to hardware and system software 20%
Perfective User enhancement, improved documentation, recoding for computational efficiency 60%

3.5 Maintainable designs (p.213). To reduce the need for maintenance and do essential tasks efficiently: (a) define user requirements more accurately during development; (b) make better system documentation; (c) use proper methods of designing processing logic and communicating it to the team; (d) use existing tools and techniques effectively; (e) manage the systems engineering process better. "Design is both a process and a product."

3.6 Testing practice and plans (pp.213–214). The philosophy is to find errors. A test case is "a set of data that the system will process as normal input" but created with the express intent of finding whether the system processes it correctly (inventory example: quantities to withdraw that exceed, equal and are less than the quantities on hand). Two general strategies:

Strategy What the analyst does
Code testing Develops cases to execute every instruction and path in a program.
Specification testing Examines the program specifications and writes test data to see how the program operates under specific conditions.

3.7 Levels of tests (pp.214–216). "The analyst must perform both unit and system testing."

Unit testing (3.7.1, p.214) System testing (3.7.2, pp.214–216)
Scope The programs/modules making up a system, tested independently of one another; also called program testing All subsystems gathered into one pool and tested as a whole
Finds Errors in coding and logic contained within that module alone; interaction errors are initially avoided Weaknesses not found in earlier tests; "the last chance to detect and correct errors before the system is installed for user acceptance testing"
Method Test cases exercise each condition and option; can be done bottom-up from the smallest, lowest-level modules, using a short driver program per module Considers all likely variations and pushes the system to its limits
Example Hotel system: reservations, check-in/check-out, restaurant and room-service charges, convention activities, accounts-receivable billing Low volumes of live transactions, raised to the maximum for each transaction type, with the old system still running

The number and nature of errors depend on communication between user and designer, the programmer's ability to code the specification exactly, and the time frame. System testing consists of five steps (p.215): program testing, string testing, system testing, system documentation, user acceptance testing.

Step Course content
Program testing (p.215) Program must compile and test data correctly and tie in with other programs; the programmer is responsible for an error-free program. Two kinds of error: a syntax error violates a language rule (improperly defined field dimension, omitted keywords; shown by compiler error messages); a logic error concerns incorrect data fields, out-of-range items and invalid combinations; not detected by the compiler, so the programmer examines the output. Actual output is compared with expected output; on a discrepancy the instruction sequence is traced (p.216).
String testing (p.216) Programs interact; each is tested to see whether it conforms to related programs, using test and live data, before the whole system is tested.
System testing (p.216) Uncovers weaknesses not found earlier, including forced system failure and validation of the total system in the operational environment; low volumes of live-data transactions are increased to the maximum per type; tests recovery and fall-back after major failures so no data is lost; done with the old system still in operation; the old system is discontinued only when the proposed one succeeds.
System documentation (p.216) All design and test documentation prepared and kept in the library, the central location for maintenance of the new system.
User acceptance testing (p.216) Objective: "selling the user on the validity and reliability of the system"; verifies that procedures operate to specification and the integrity of important data is maintained; "the user's show"; a comprehensive test report shows the system's tolerance, performance range, error rate and accuracy.

3.8 Special systems tests (p.217). "There are other six tests":

# Test What it determines Course example
(i) Peak load test whether the system handles the volume at peak processing demand activate all terminals at the same time
(ii) Storage testing capacity to store transaction data on disk or in files verify the system stores 10,000 records of 400 bytes on a single flexible disk
(iii) Performance time testing length of time to process transaction data, before implementation time to get a response to an inquiry, copy a file, send a transmission and get a response
(iv) Recovery test ability to recover data or restart after failure load a backup copy and resume without data or integrity loss
(v) Procedure testing clarity of operating and use documentation, by having users do exactly what the manuals say power down at week end; respond to the printer's paper-out light
(vi) Human factors testing how users will use the system when processing data or preparing reports (none given)

3.9 Designing test data (p.218). Poor test data makes the output unreliable.

Live data Artificial data
Source Actually extracted from users' files; users key in a set from normal activities after the system is partly built Created solely for test purposes
Property Biased towards typical values; hard to obtain in sufficient amount Provides all combinations of values and formats; tests all logic and control paths; provides extreme values for testing limits

3.10 System control (pp.218–219). Controls mean "the extent to which the system is secure against human errors, machine malfunction or deliberate mischief"; the amount varies with importance, criticality and volume of output. Objectives (p.218–219):

Objective Meaning
(a) Accuracy information and reports accurate in all respects
(b) Reliability the system keeps functioning as designed, no breakdown from equipment malfunction
(c) Security files and programs secured against accidental damage or loss; access restricted to authorised users
(d) Efficiency tasks done in the best manner, highest quality output with least effort
(e) Audit a transaction can be traced from creation to final use (especially online)
(f) Adherence to organisation policies controls do not conflict with, and promote, organisation policy (example: payroll accurate and on time)

Too many controls hurt performance; too few cause chaos: a trade-off by the nature and criticality of the system. Types of control (p.219):

External control Internal control
Meaning laws, regulations, procedures and policies outside the scope of the system that affect its operation plans, procedures, guidelines, rules and checks under which the system must function
Source outside much specified in system design; some are the organisation's own (with or without computerisation)

3.11 Audit trail (p.220). In online systems there may be no copies of source documents to fall back on; users can sign on, change stored data and sign off with no visible clue. Audit trail is "the path which a transaction traces through a data processing system from source documents to summary reports": the facilities or procedures that allow a transaction to be traced from its appearance on a source document to its transformation into information on a final output. It holds complete details (reference numbers, dates, names) in files, ledgers and journals so records can be traced back to source documents. Generally available in manual accounting; in computerised systems not available unless specially designed. Audit trails are "not primarily for the use of auditors": they are tools for management, which auditors then use.

Code and worked examples

No code is used in this unit; examples are test designs and classifications.

Example 1 — Test cases for inventory withdrawal (Course example pp.213–214). The course says test quantities should exceed, equal and be less than the quantity on hand. Take on-hand = 50 (illustrative; the expected results are a sensible rule, not a course statement).

Case Quantity to withdraw Relation to on-hand Expected result
1 60 exceeds rejected with an error message
2 50 equals accepted, on-hand becomes 0
3 40 less than accepted, on-hand becomes 10

A test is successful when it finds an error: if case 1 were accepted, the test has succeeded in showing a defect.

Example 2 — Classify maintenance (Table 2, p.213).

Scenario Type Reason
Emergency fix after a payroll crash Corrective emergency fixes, routine debugging
Update the program because the printer was replaced Adaptive accommodates changes to hardware/system software
Users ask for a new report column Perfective user enhancement
Rewrite a loop to run faster Perfective recoding for computational efficiency
Improve the user manual Perfective improved documentation
File layout changed by the bank Adaptive changes to data and files

Arithmetic with the course's frequencies on 500 maintenance requests: corrective 20% of 500 = 100; adaptive 20% = 100; perfective 60% = 300. Perfective alone (300) exceeds corrective plus adaptive together (200), as Unit 4 p.244 also states.

Example 3 — Reliability approaches (Table 1, p.211).

Situation Approach
Program is written so illegal input cannot occur Error avoidance (the course notes this is impossible in large systems)
Program traps a divide-by-zero and substitutes a safe value Error detection and correction
Printing module fails; the system disables printing but continues processing orders Error tolerance (degraded performance)

Example 4 — Order of the five system-testing steps (p.215) on a hotel system (p.214).

Order Step Hotel-system activity
1 Program testing Reservation module compiled, syntax and logic errors fixed against test data
2 String testing Reservation output passed to check-in and billing programs; interface checked
3 System testing All modules together, transaction volume raised from low to peak; recovery after forced failure tested; old system still running
4 System documentation Design and test documents filed in the library
5 User acceptance testing Hotel staff run it; test report shows tolerance, performance range, error rate, accuracy

Unit testing (bottom-up) happens within step 1: each module is run alone, with a small driver program feeding it data.

Example 5 — Audit trail trace (definition p.220; the sale is illustrative). A sales order moves through the system:

Text
Source document: Order form #SO-0412, 12 Nov, customer C-77        (reference no., date, name)
   -> Entry record in sales file: SO-0412
   -> Journal/ledger posting: sales ledger entry J-3391 refers to SO-0412
   -> Summary report: "Sales by customer, November" line for C-77 includes J-3391

Because each hop carries the reference number and date, a query on the report line can be traced back through J-3391 and SO-0412 to the source form. In an online system with no audit trail, a user could change the stored balance of C-77 and sign off leaving no clue.

Example 6 — Verification vs validation (pp.210).

Verification (alpha) Validation (beta)
Environment simulated live, at selected user (beta) sites
Transactions test live, day-to-day
Duration before release may continue for several months
Outcome feedback changes the software failures found in use change the software again

Common mistakes

  • Verification vs validation. Verification = simulated environment (alpha testing); validation = live environment (beta test sites). An MCQ may reverse them.
  • Unit testing vs system testing. Unit testing = one module at a time, also called program testing; system testing = all subsystems together. "Program testing" is also the first of the five system-testing steps.
  • Successful test. One that finds an error, not one that shows the program is error-free.
  • Syntax vs logic errors. Syntax errors are caught by the compiler; logic errors are not.
  • Maintenance types. Corrective 20%, adaptive 20%, perfective 60%; perfective is the biggest (user enhancement). Do not call adaptive "adoption".
  • Internal vs external control. External = laws, regulations, outside policies; internal = plans, procedures, rules and checks within the system. Audit trails are for management, not primarily for auditors.
  • Test data. Live data is biased to typical values; artificial data supplies extreme values.

Errata

Page Course says Correct Note
217 Special tests numbered "(i), (ii), (iii), (vi) Recovery Test, (v), (vi)" Recovery test is (iv) Two items numbered (vi); the course text says there are six tests.
213 Table 2, perfective: "recording for computational efficiency" recoding (rewriting code) for computational efficiency Typographical slip.
217 "Recovery Test ... ability of user to recover data or re-start system" Ability of the system to recover/restart Wording slip; the example (load a backup and resume) shows the test is of recovery after failure.

Practice questions

  1. (MCQ) According to the course, a successful test is one that: A. finds no errors B. finds an error C. runs faster than before D. is accepted by the user
  2. (MCQ) Using software in a live environment to find errors is called: A. verification B. validation C. certification D. benchmarking
  3. (MCQ) The three levels of quality assurance are: A. unit, system, acceptance B. testing, verification and validation, certification C. code, specification, string D. peak load, storage, recovery
  4. (MCQ) A system that recognises an error but keeps running with degraded performance (for example by shutting down part of the system) uses: A. error avoidance B. error detection and correction C. error tolerance D. error prevention
  5. (MCQ) Which maintenance type has the highest relative frequency (60%) in Table 2? A. Corrective B. Adaptive C. Perfective D. Preventive
  6. (Fill) The five steps of system testing are program testing, ______ testing, system testing, system ______ and user ______ testing.
  7. (MCQ) Which error is not detected by the compiler? A. Syntax error B. Omitted keyword C. Logic error D. Improperly defined field dimension
  8. (MCQ) Activating all terminals at the same time is the course's example of: A. storage testing B. peak load test C. recovery test D. procedure testing
  9. (Self-assessment 1 and 2) (a) List the factors responsible for the quality of a system. (b) Explain the levels of quality assurance. (c) Explain the importance of system reliability. (d) Name the types of maintenance. (e) Describe unit testing. (f) Name the special system tests.
  10. (TMA) (a) List the objectives of system control. (b) Difference between internal and external control. (c) What is an audit trail?
Answersshow
  1. B: "A successful test, then, is one that finds an error" (p.210).
  2. B: validation = live environment; verification = simulated (p.210).
  3. B: p.210.
  4. C: error tolerance, Table 1 (p.211).
  5. C: perfective 60% (p.213); the course has no "preventive" type.
  6. string; documentation; acceptance (p.215).
  7. C: logic errors are not detected by the compiler (p.215).
  8. B: peak load test (p.217).
  9. (a) Correctness, reliability, efficiency, usability, maintainability, testability, portability, accuracy, error tolerance, expandability, access control and audit, communicativeness (pp.208–209). (b) Testing (executing a program to find errors); verification (simulated environment, alpha) with validation (live environment, beta sites); certification (package conforms to standards and vendor's claims; not necessarily the best package) (p.210). (c) A reliable system does not produce dangerous or costly failures in normal use; reliability has two levels (right requirements; actual working) and three approaches (error avoidance, detection and correction, tolerance) (p.211). (d) Corrective, adaptive, perfective (p.213). (e) Testing the programs/modules making up a system independently, also called program testing, to find coding and logic errors within each module; can be bottom-up with a short driver program (p.214). (f) Peak load, storage, performance time, recovery, procedure, human factors (p.217).
  10. (a) Accuracy, reliability, security, efficiency, audit, adherence to organisation policies (pp.218–219). (b) External = laws, regulations, procedures and policies outside the system that affect it; internal = plans, procedures, guidelines, rules and checks under which the system must function (p.219). (c) The path a transaction traces from source document to summary report, with reference numbers, dates and names, so records can be traced back; a management tool (p.220).

Module 3 · Unit 3 of 4

Unit 3 — Documentation

What you must know

  • List the five characteristics of good documentation.
  • Name the five types of documentation and say who each is written for and what it contains.
  • Describe the three documentation tools: structured (Nassi–Shneiderman) flowchart, HIPO diagram, Warnier/Orr diagram.
  • Give the eight purposes of documentation and the three factors that shape a documentation package's format.
  • List the eight elements of a documentation package and the seven items of a file specification.

Key ideas

Intuition. Documentation is the system's memory. People leave; the system must still be understood and changed. Each reader (programmer, operator, user, manager, the next analyst) needs a different document.

Introduction (p.222). "A system cannot be completely effective unless it is adequately documented." It should be documented as it is being created: status reports at each stage (flowcharts, decision tables, output forms, problems found, suggested solutions, schedule revisions) keep management aware so changes can be made while still economically and physically possible. These progress reports are a basis for further documentation.

3.1 Characteristics of good documentation (p.223). Five qualities:

Quality Meaning
(a) Availability accessible to those for whom it is intended
(b) Objectivity clearly defined in a language easily understood
(c) Cross-referable possible to refer to other documents
(d) Easy to maintain easy to update when the system is modified
(e) Completeness contains everything needed for a careful reader to understand the system

3.2 Types of documentation (p.223): five major types.

Type For whom Content (course)
3.2.1 Program documentation (pp.223–224) Programmers (present and future) Analysts give the programmer documentation before the program is developed; logic is best shown by a flowchart or a decision table. Four items constitute normal documentation for each program: (1) copies, in final form, of all input/output documents affecting the program; (2) a statement of standards for coding structures and input/output layouts; (3) clarification of the program's interface with other related programs; (4) a general flowchart or decision table. Purpose: let future programmers make changes (programmer turnover is high). A documentation folder per program holds the I/O forms, a detailed flowchart or decision table and a set of operator and user instructions. Costly and time-consuming, but the changes that require updating are the very reason for accurate documentation.
3.2.2 Operations documentation (pp.224–225) Console operators The operator must know the job completely, in a form readily accessible and in simple language; the analyst discusses requirements of new jobs with operations staff before transfer. The run book (traditional; a loose-leaf notebook so sheets can be replaced) contains: narrative describing the run; listing of programmed error conditions; detailed running information: I/O forms to be used, anticipated problem areas and how to handle them, file assignment of each I/O device, disposition of data files after the job, general block diagram of the programming logic, restart procedures.
3.2.3 User documentation (pp.225–226) Users and their supervisors A manual that says everything users must know to do their jobs. Users need two kinds of information: complete details for each case the system processes, and the overall picture of their role in the whole. Manual contents (seven): general flowchart of the system; assignment of responsibility for tasks; standards for work flow (target dates and deadlines); simple input and output documents; detailed procedures; anticipated exceptions and how to handle them; accuracy standards for data. The systems department writes a detailed narrative of each system including routine cases and exceptions; supervisors need charts, graphs and illustrations of their department's role.
3.2.4 Management documentation (p.226) Corporate management Enables management to (i) evaluate progress on systems development, (ii) monitor existing systems, (iii) understand objectives and methods of new and existing systems. Management wants the overall objectives and basic operations: a brief manual of the key steps; summaries and charts.
3.2.5 Systems documentation (pp.226–227) Analysts, designers, selection committee Each phase of the development cycle has documentation. The systems request (written, ideally jointly by client and analyst) is a statement of the user's problem. The selection committee documents: (a) objectives of the feasibility study, (b) extent of the authority of the feasibility team, (c) who is responsible for completing the study. The feasibility report is "probably the most important form of documentation in the system development cycle": it (i) defines the objectives of the proposed change in reasonable detail, (ii) gives a plan to attain them. Design adds: file specification (detailed definitions of each file, best done graphically), transaction specification and output specification. Also plans to test, to convert from the old system and to train personnel. The finished system should yield data about its own operation as a by-product.

3.3 Software design and documentation tools (pp.227–230). "Well designed, modular software is more likely to meet the maintenance, reliability and testing requirement." Three tools:

Structured flowchart (3.3.1, p.228) HIPO diagram (3.3.2, pp.228–230) Warnier/Orr diagram (3.3.3, p.230)
Other name Nassi–Shneiderman chart Hierarchical Input Process Output logical construction of programs / systems
Idea Graphic tool that forces the designer to structure software in modular, top-down form Highly structured modular design, drawn like an organisation chart; top-to-bottom approach; data flows down through the system Designs program structure by identifying the output and processing results, then working backwards to the steps and input needed
Elements Process (rectangular box: initialisation, I/O, calls to other procedures), Decision (alternative conditions, can show more than two alternatives), Iteration (looping while or until a condition) Numbered boxes identify every sub-level and component (Fig. 1: 0; 1, 2, 3; 11, 12, 13; 21, 22; ... down to five digits) Brackets group the components of each process; brackets denote sets and subsets (Fig. 2)
Rules No arrows and no continuations on separate pages; one sheet; logic specified top-down; single exit Aim: eliminate "output-oriented" system solutions, which are fragmentary, have logic gaps and duplicate programs Shows processes and their sequences hierarchically; simple appearance, easy to understand
Original purpose Gives a structure the programmer keeps for developing the application software Provide documentation help to programmers/analysts Show levels and data movement between them

(An output-oriented system "is concerned with providing output and does not bother about the sound principles of system design"; it "gets the job done without delay" but wastes effort, p.229.)

3.4 Need for documentation (p.231). It depicts what the system should be and how it should perform; shows technically and economically how it serves the company; improves overall operation, management and audit control. Eight purposes:

  • (i) reviews the progress or development of an application;
  • (ii) communicates facts about the system to users;
  • (iii) communicates between personnel on a development project;
  • (iv) gives guidelines to correct or revise a system or its programs;
  • (v) provides operating instruction to users and operating staff;
  • (vi) helps reconstruct the system if it is destroyed;
  • (vii) helps management decide whether the new design achieves company objectives within constraints and is cost-justified;
  • (viii) is a focal point from which the analyst's design is assessed and a standard for reference after implementation.

3.5 Guidelines/format (pp.231–232). "There are, as yet, no universal documentation standards." The format of each package depends on: (i) characteristics of the system (descriptive vs diagrams); (ii) management's attitude towards documentation (prepare within the limits management sets); (iii) equipment restraints (large integrated teleprocessing systems need more formal and technical documentation than small, conservative ones).

3.6 Elements of a documentation package (pp.232–233). Eight elements:

# Element Content
(i) Cover letter Correspondence to management describing benefits and helping to sell the system; concise; states purpose and function clearly. Unless the documentation is approved the system will never be implemented.
(ii) Table of contents An "absolute necessity"; pages numbered and cross-referenced.
(iii) Narrative Begins the detailed formulation of the new system.
(iv) Flowcharts Each subsystem of the formal design explained with a flowchart.
(v) File specification Each file: purpose; programs that use it; volume; frequency of use; source; description of fields; layout and samples.
(vi) Program specification Segment the design so each unit has a separate program (once the design is approved).
(vii) Cost of the proposed system and of its alternatives Cost details shown.
(viii) Test brochures Operations and procedures, including test data, for testing the new system once approved.

Summary (p.234). Documentation takes place at every stage; some must be updated at project end and during system reviews.

Code and worked examples

Example 1 — Matching documentation to the reader (types, pp.223–227).

Item Document type Reason
"Restart procedures for the nightly job" Operations (run book) operator instruction
"Deadlines and accuracy standards for data entry clerks" User documentation user manual contents
"Brief summary of key steps and objectives for the managing director" Management documentation evaluate, monitor, understand
"Flowchart and I/O layouts for the payroll program" Program documentation one of four program items
"Feasibility report and file specification" Systems documentation produced during the development cycle

Example 2 — Warnier/Orr set notation (Course example, Fig. 2, p.230). The course shows the days of the week as nested brackets; redrawn as text:

Text
                         ( Monday
            ( Week days  ( Tuesday
Days of the  (           ( Wednesday
week         (           ( Thursday
            (            ( Friday
            ( Week end   ( Saturday
              days       ( Sunday

Reading: the set "Days of the week" has two subsets, "Week days" (five members) and "Week end days" (two members). Each bracket groups the components of one level, which is how Warnier/Orr diagrams show hierarchy. (Warnier/Orr diagrams work backwards from the output.)

Example 3 — HIPO numbering (Course example, Fig. 1, p.229). The chart is a tree where each box's number identifies its position; redrawn from the figure as a text outline (top levels only):

Text
0
|-- 1
|   |-- 11  (111, 112, 113 ...)
|   |-- 12
|   `-- 13  (131 -> 1311, 1312)
|-- 2
|   |-- 21  (211, 212)
|   `-- 22  (221, 222)
`-- 3
    |-- 31  (311, 312)
    `-- 32  (321, 322)

Rule of reading: the first digit is the top-level branch; each added digit is a sub-component of the box whose number it extends (e.g. 131 belongs to 13, which belongs to 1). Data flows down the tree.

Example 4 — The if-then of Unit 1 as a structured (Nassi–Shneiderman) chart (concept pp.228). Structured flowcharts use no arrows; the same logic as Unit 1 Fig. 4 and a loop (illustrative layout):

Text
+----------------------------------------------+
| Read an employee record                      |
+----------------------------------------------+
| Iteration: WHILE there are more records      |
|   +----------------------+-----------------+ |
|   | Decision: years > 10?                  | |
|   |   YES                |     NO          | |
|   | Print employee name  |  (no action)    | |
|   +----------------------+-----------------+ |
|   | Read next employee record              | |
+----------------------------------------------+

It has one entry at the top, one exit at the bottom and no arrows or page continuations; process, decision and iteration are the only element types the course lists.

Example 5 — Assembling a documentation package (p.232–233). Put these in the course's order: Test brochures; Narrative; Cover letter; Cost of proposed system and alternatives; Table of contents; File specification; Flowcharts; Program specification.

Order: 1 Cover letter; 2 Table of contents; 3 Narrative; 4 Flowcharts; 5 File specification; 6 Program specification; 7 Cost of the proposed system and of its alternatives; 8 Test brochures. For element 5, a file specification for a customer file must give: purpose (hold customer details), programs using it (billing, statements), volume, frequency of use, source, description of fields, layout and samples: seven items.

Common mistakes

  • Five types, not four: program, operations (operation), user, management, systems documentation.
  • Program documentation has four items; user manual has seven; run book has three parts (narrative, error conditions, detailed running info); package has eight elements; need for documentation has eight purposes. MCQs often swap counts.
  • Structured flowchart vs HIPO vs Warnier/Orr. Structured flowchart = Nassi–Shneiderman, no arrows, three elements; HIPO = hierarchy, numbered boxes, top-to-bottom, like an org chart; Warnier/Orr = brackets, works backwards from output.
  • Output-oriented systems are what HIPO tries to eliminate, not what it builds.
  • Cover letter is aimed at management to sell the system; the table of contents is "an absolute necessity".
  • Feasibility report (not the systems request) is "probably the most important" documentation in the cycle.

Errata

Page Course says Correct Note
227 "(ii) Transaction Specification: detailed descriptions of all output anticipated from the system. (iii) Output Specifications: detailed descriptions of all output anticipated from the system." Transaction specification should describe the transactions (inputs) the system processes Same sentence printed twice; the first is evidently a copying slip. For the exam, know that the three are file, transaction and output specification.
228 "Nassi-Schneiderman charts" Nassi–Shneiderman Spelling of the inventors' names (Isaac Nassi and Ben Shneiderman). Checked at the external page listed below.
224 vs 223 Heading 3.2.2 "Operations Documentation"; list at p.223 "Operation Documentation" Same item Naming difference only.

Practice questions

  1. (MCQ) How many major types of documentation does the course list? A. 3 B. 4 C. 5 D. 6
  2. (MCQ) Which is not one of the five characteristics of good documentation? A. Availability B. Cross-referable C. Completeness D. Confidentiality
  3. (MCQ) The run book is documentation for: A. users B. console operators C. management D. auditors
  4. (MCQ) Which of these is not one of the four items normally required for program documentation? A. Input/output documents in final form B. Coding and layout standards C. Program interface with other programs D. A cover letter to management
  5. (MCQ) The three functions management documentation lets management perform are evaluate progress, monitor existing systems and: A. write programs B. understand objectives and methods of new and existing systems C. train operators D. prepare run books
  6. (MCQ) HIPO stands for: A. Hierarchical Input Process Output B. Hierarchy Input Program Output C. Hardware Input Process Operation D. High Input Processing Output
  7. (MCQ) Which diagram uses brackets and works backwards from the output? A. HIPO B. Structured flowchart C. Warnier/Orr D. Run book
  8. (Fill) A structured flowchart uses no ______ and its three element types are process, decision and ______.
  9. (Self-assessment 1 and 2) (a) Describe the characteristics of good documentation. (b) List the types of documentation. (c) What is user documentation? (d) List the software design and documentation tools. (e) Explain the HIPO diagram. (f) Explain the Warnier/Orr diagram.
  10. (TMA) (a) Why is documentation needed? (b) What elements comprise a documentation package?
Answersshow
  1. C: five (p.223).
  2. D: availability, objectivity, cross-referable, easy to maintain, completeness (p.223).
  3. B: operations documentation (pp.224–225).
  4. D: the four items are I/O documents, coding/layout standards, interface with other programs, general flowchart or decision table (p.224).
  5. B: p.226.
  6. A: p.228.
  7. C: Warnier/Orr (p.230).
  8. arrows; iteration (p.228).
  9. (a) Availability, objectivity, cross-referable, easy to maintain, completeness (p.223). (b) Program, operations, user, management, systems (p.223). (c) A manual that spells out everything users must know to do their job: general flowchart, responsibilities, work-flow standards, simple I/O documents, detailed procedures, anticipated exceptions, accuracy standards (p.225). (d) Structured flowchart, HIPO diagrams, Warnier/Orr diagrams (p.227). (e) Hierarchical Input Process Output: modular, organisation-chart structure, numbered boxes, top-to-bottom, data flowing down, aimed at eliminating output-oriented design (pp.228–229). (f) Also called logical construction of programs/systems; designs program structure by starting from the output and working back to the inputs; brackets show sets and subsets and the levels (p.230).
  10. (a) Eight purposes: review progress; communicate facts to users; communicate among project staff; guide correction or revision; give operating instruction; help reconstruct a destroyed system; help management judge the design against objectives and cost; focal point and standard for assessment and later reference (p.231). (b) Cover letter, table of contents, narrative, flowcharts, file specification, program specification, cost of the proposed system and its alternatives, test brochures (p.232).

Module 3 · Unit 4 of 4

Unit 4 — System Implementation

What you must know

  • Name the three aspects of implementation and describe operator and user training and the training methods.
  • Compare the four conversion methods (parallel, direct, pilot, phase-in) with their advantages and disadvantages.
  • Describe the conversion plan: tasks, site preparation, and file and data conversion controls (record counts, financial controls, hash totals).
  • Explain the post-implementation review and the four parts of the review plan; state maintenance facts and types.
  • List the contract clauses, hardware acquisition factors, vendor selection criteria and proprietary software evaluation approaches.
  • Compare service bureaux with in-house computing and compare renting, leasing and outright purchase.

Key ideas

Intuition. A good system fails if it is installed badly. Implementation is: train the people, switch from the old system safely, then review and maintain it, and buy hardware, software and services sensibly.

Introduction (p.236). "Implementation includes all those activities that take place to convert from the old system to the new one." The new system may be completely new (replacing a manual or automated system) or a major modification. "Successful implementation may not guarantee improvement in the organization using the new system, but improper installation will prevent it." The unit discusses three aspects: training personnel, conversion procedures, post-implementation review.

3.1 Training of personnel (pp.237–238). The quality of training affects success; those involved must know their roles, how to use the system efficiently and what it will or will not do for them. Both system operators and users need training.

System operators (3.1.1, pp.237–238) Users (3.1.2, p.238)
Who Computer-centre staff who support the system People who enter data and use output
Content Handle all operations, routine and extraordinary; for new equipment: how to turn it on, use it, power off, what is normal operation; recognise malfunctions and the steps to take; a troubleshooting list (problems and remedies) and names and telephone numbers of contacts; familiarisation with run procedures (the sequence of activities to use the system on an ongoing basis) Equipment use where the person is both operator and user (turn on a terminal, insert a diskette, when it is safe to power off); mainly operation of the system itself and data handling: entering transactions, editing data, formulating inquiries, deleting and inserting records; plus simple system maintenance activities
Risk Weakness in any aspect leads to user frustration and errors

3.2 Training methods (pp.238–239). Two most important ways:

Vendor and in-service training (3.2.1) In-house training (3.2.2)
Idea Often the best source for equipment training; vendors offer extensive programmes (course example: IBM offers complimentary two or three days of short courses to purchasers of many minis and mainframes, one session hands-on). Special software (teleprocessing package, DBMS): off-site short-term, in-depth courses, usually for a fee, with personnel from many organisations Run inside the organisation
Advantages Experienced trainers; sharing questions, problems and experience with people from other companies is "substantial" Instruction tailored to the organisation's requirements; vendors often negotiate more economical fees, so more personnel can be trained than when travel is required
Disadvantages Fees; travel Distractions: telephone calls, business emergencies, interruptions

A common approach to evaluate in-house training: a case-study example containing all frequently encountered situations; the user enters data, processes it and prepares the reports. "High-quality training is an essential step in systems implementation, yet it is not sufficient by itself" (p.239).

3.3 Conversion methods (pp.239–240). "Conversion is the process of changing from the old system to the new one." Four common methods: parallel systems, direct conversion, pilot system, systems phase-in. Sometimes one may be forced to use a method even if another would be better. Generally conversion should take the shortest possible time; long conversions cause problems for all involved.

Parallel Direct Pilot Phase-in
Idea Run both systems; users operate the old system as usual and also start using the new Convert abruptly, sometimes over a weekend or overnight; the old system is used until conversion day then replaced A working version is installed in one part of the organisation (a work area or department); users know it is a pilot and changes can be made Used when the system cannot be installed everywhere at once: staged over weeks to months (file conversion, training or equipment arrival force it)
Advantages "The most secure" and "safest": if the new system has problems the organisation can fall back to the old without loss of time and money No parallel activity; full reliance on the new system Experience and live test before full implementation; preferred for new techniques or drastic changes; feedback improves the system before the rest of the organisation (all at once = direct, or gradually = phase-in) Some users benefit early; training and installation without unnecessary use of resources
Disadvantages Doubles operating costs; the new system may not get a fair trial No other system to fall back on; requires wise and careful planning (the course lists none; it takes time before the whole organisation is converted) (the course lists none)

(Parallel systems are on pp.239–240; direct conversion, pilot system and phase-in on p.240.)

3.4 Conversion and operation plans (pp.240–242). The conversion plan defines all activities needed to implement the new system, identifies who is responsible for each and gives a time schedule (p.240). Pre-implementation task list (p.241): list all files for conversion; identify all data needed to build new files; list all new documents and procedures used during conversion; identify all controls to be made during conversion; assign responsibility for each task; verify conversion schedules. The plan should anticipate problems: missing documents, mixed data formats between current and new files, errors in data translation, situations overlooked in development. The conversion manager guards against omitted steps, reviews plans, verifies delivery of equipment and software and prepares the site.

  • 3.4.1 Site preparation (p.241). Electrical and air-conditioning preparation, site layout, equipment installation; best completed before equipment arrives (vendors will not deliver while construction goes on). For a microcomputer little work is needed, but check power lines for static or fluctuations and use a "clean" line not shared with other equipment; avoid carpets (or use anti-static carpet). The layout leaves space for moving equipment in, service and maintenance, and air circulation; vendors' clearance requirements must be followed or warranties can be voided and maintenance discontinued.
  • 3.4.2 File and data conversion (p.242). Data and file preparation consumes a large proportion of conversion time. Data must be converted to the new format without loss of detail or accuracy. Using record counts, financial controls and hash totals, analysts detect and correct problems quickly, even when conversion involves data transmission.

3.5 Post-implementation review (p.242). After implementation and conversion, a review determines whether the system meets expectations and where improvements are needed. It "measures the system's performance against pre-defined requirements", determines how well it continues to meet performance specifications, and shows whether major re-design or modification is required. It evaluates the extent to which the system accomplishes its stated objectives and how actual project costs compare with initial estimates (the printed sentence is garbled), and reviews major problems still needing correction and those that surfaced in implementation. The study begins with a review team that gathers and reviews requests for evaluation; an unexpected change that affects users or performance is a primary trigger; the user is asked how well the system functions to specification, how far measured benefits are realised and for suggestions for changes.

3.6 Review plan (pp.242–244). The review team prepares a formal plan around the objectives, the type of evaluation and the time schedule. Four parts:

Plan Reviews
(i) Administration (administrative) plan User objectives (the system may no longer meet objectives, or objectives may have changed; results documented) and operating costs and benefits (current budget against costs and savings)
(ii) Personnel requirements plan All activities involving system personnel: performance objectives compared with current performance; training performance, through testing, interviews and other data gathering
(iii) Hardware plan Terminals, CRT, communication network; compare current performance specifications with design specifications; point out needed modifications
(iv) Documentation review plan Accuracy and completeness of documentation to date and conformity with documentation standards set earlier

The review is also a source of information for the next projects.

3.7 System maintenance (pp.244–245). "The last part of the system development life cycle is system maintenance which is actually the implementation of the post-implementation review plan." Average life of a system 4 to 6 years, oldest often over 10. Findings: (a) 60–90% of total software cost over a system's life is maintenance; (b) in documented cases the cost of maintenance, measured per instruction written, is more than 50 times the cost of developing the system; (c) software demand grows faster than supply; more programmer time goes to maintenance than new development; there is a backlog of new work. Maintenance is classified as corrective, adaptive (printed "adoptions") or perfective:

  • Corrective: repairing processing or performance failures, or alterations because of previously ill-defined problems.
  • Adaptive ("adoption" in the text): changing program functions.
  • Perfective: enhancing performance or modifying programs for users' additional or changing needs; the greatest amount of work (user enhancement and improved documentation). "More time and money are spent on perfective than on corrective and adaptive maintenance together."

Maintenance covers correcting coding and design errors, updating documentation and test data and upgrading user support. "Maintenance means restoring something to its original position. Unlike hardware, software does not wear out; it is corrected." Enhancement means adding, modifying or re-developing code to support changes in specifications. To reduce the need for maintenance (p.245): define requirements more accurately; better documentation; more effective ways to design processing logic and communicate it; better use of tools and techniques; manage the systems engineering process effectively. The maintenance programmer's environment also matters (less paid, less recognition, little training; maintenance demands more orientation and training than other programming).

3.8 Drawing up a computer contract (pp.245–247). Vendors usually offer their own standard contract, drafted in the vendor's interest; the buyer can negotiate to add, delete or modify clauses. Buyer's safeguards:

Clause Content
3.8.1 Responsibilities of vendor and buyer (p.246) Responsibilities and remedies for non- or faulty performance. Remedies: special remedies (invoked for late delivery or deficiency at acceptance tests); damages: actual (compensate for advantages lost in the actual bargain), consequential (compensate all foreseeable losses from breach), liquidated; and specific performance (applies when the system fails acceptance tests because of memory storage: buyer can ask for additional memory at nil or nominal cost within a specified time).
3.8.2 Documentation Customer may stop payment if the documentation needed to run the system is not provided.
3.8.3 Hardware Each component identified and its performance criteria stated under normal operating conditions.
3.8.4 Delivery and acceptance Delivery schedule and the acceptance tests/standards to be met under normal operating conditions during a specific period of use.
3.8.5 Right of use of equipment from other vendors Customer free to link peripherals from other manufacturers to the CPU; the vendor is protected from damage or extra costs resulting.
3.8.6 Warranties Should be included.
3.8.7 Guarantees (p.247) For reliability: minimum hours of usable time per day; mean-time-between-failures (MTBF); maximum time to repair.
3.8.8 Payments Whether rental/lease payments are by hours used per month/shift or a flat monthly rate; lease contracts include a buy option at stipulated prices; protect negotiated prices; if the vendor insists on escalation clauses the customer may terminate and recover preparation costs.
3.8.9 Bankruptcy Explicit protection if the vendor becomes bankrupt.

3.9 Hardware acquisitions (pp.247–249). Once in-house computing is decided, list specifications so vendors can be invited. Tender specifications follow the approved feasibility report, include main technical parameters, project objectives and the implementation schedule; vendors may quote separately for leasing and buying and give details of the infrastructure the customer must arrange and its likely cost.

Section Content
3.9.1 Tender evaluations (p.248) Offers far below essential requirements are rejected from short-listing; marginal shortfalls considered on merit; "desirable" additional features get appropriate weights so all bids are on an equal footing. Quantifiable differences: one-time costs and continuing running and maintenance costs; equipment characteristics (storage capacity, speeds); built-in spare capacity and ability to support additional peripherals; additional vendor support.
3.9.2 Costing factor Two types: (i) one-time costs (site preparation: space, false ceilings, special flooring, electrical fittings, air-conditioning) and (ii) continued running and maintenance costs.
3.9.3 Equipment characteristics Higher transfer rates, larger storage capacity, for printers characters per line; high MTBF; compatibility with other equipment and peripherals.
3.9.4 Potential for growth (pp.248–249) Growth beyond current capacity by adding components; growth within a family of computer models; handling a large variety of peripherals; handling additional workloads considering peak hour loads.
3.9.5 Vendor support (p.249) Hardware maintenance facilities; training; help with software development; back-up in case of failure; comparative delivery periods.

3.10 Criteria for vendor selection (pp.249–250). Mandatory requirements: a vendor that fails them is screened out. Desirable characteristics are harder to evaluate. Criteria, "in descending order by importance":

Group Items
3.10.1 Economic factors cost comparisons; return on investment; acquisition method
3.10.2 Hardware factors performance and reliability; facilities for modularity; provision for back-up; firmness of delivery date; compatibility with existing systems
3.10.3 Software factors performance and price; efficiency and reliability of available storage; programming languages available; availability of well-documented package software; firmness of software delivery date; ease of use and modification per user requirements; portability and capacity to interface with the environment
3.10.4 Service factors facilities for detecting errors in new programs; good training facilities; assistance in software development and conversion; maintenance terms and quality
3.10.5 Reputation of manufacturer financial stability; past history of keeping promises

3.11 Acquiring proprietary software packages (pp.250–252). Proprietary rights usually stay with the original developer, so a package can only be licensed. Payment modes: (1) off-the-shelf programs on floppy or plug-in ROM (popular for PCs); (2) a monthly/periodical licence fee; (3) one-off payment for a fixed period; (4) lump-sum at the start and periodical service charges later. Some vendors supply packages with the system for a nominal fee just to establish ownership.

3.11.1 Technical aspects (pp.251–252):

Aspect Meaning
Memory size memory required when resident in the CPU
Run time some programs use less memory but much longer run times
Adaptability fits a multiprogramming environment and uses existing peripherals, operating systems and utilities
File storage storage needed and how efficiently data is stored and retrieved
Modularity high modularity lets it run on machines with different configurations
Expandability sensitivity to handling increased volume or integrating with other programs
Reliability usable when a module fails; effect of user errors; ease of recovery; runs on different configurations
Efficiency capability under peak load; depends on language and operating system
Ease of implementation activities and vendor support until the package satisfies users and they can run it independently
Usability effort to operate, prepare input and interpret output; also portability and understandability
Vendor support documentation, modification/enhancement/maintenance and educating staff at clerical, technical, operational, managerial levels
Security built-in prevention of unauthorised access to software and data, and integrity

3.11.2 Approaches to software evaluation (p.252): (a) benchmarking (a sample program to evaluate the comparative performance of hardware or software); (b) experience of other users (ask independently, users with similar configuration and environment, not only the vendor's list of satisfied users); (c) report of independent research organisations (objective and comprehensive, published at regular intervals).

3.12 Service bureaux (pp.252–254). An in-house computer is costly; a small user may not want to justify it or set up a computer department until a few applications have shown results. Consulting firms and service bureaux provide facilities for a fee, as needed or on a continuing contract. Quality varies: some are companies renting out spare time on their own computers (low rates, negligible customer service, the user supplies operator and stationery, suits people who know EDP); others supply everything (operator, tapes, disks): dearer but customer-oriented; or the user entrusts the entire job to a data centre. Charges = fixed cost for systems development plus variable cost by volume processed.

Advantages of data centres (3.12.1, p.253) Disadvantages (3.12.2, p.254)
(i) use a computer without a large initial outlay (i) gives up control of vital business data; in-house allows strict security
(ii) no staff and management problems of employing highly paid technical professionals in a fast-changing field (ii) loses control over the time taken to process data; off-premise processing may be inconvenient
(iii) uses the expert knowledge of experienced staff (iii) own staff do not become familiar with computers
(iv) no fear of equipment becoming obsolete (iv) dependence on a second party unaware of the organisation's needs
(v) gain experience before deciding whether to install in-house

3.13 Financing the use of computers (pp.254–256). Three major approaches: renting, leasing, outright purchase; "None of the above approaches have an edge over others"; the right one depends on the organisation's characteristics and plans.

Renting Leasing Outright purchase
Term Short-term, generally 1 to 12 months, paid monthly; either side may cancel with notice (usually 30 or 60 days) A commitment for a specific time, generally three to seven years; fixed payments (monthly, quarterly, semi-annual or annual) that include equipment service and maintenance; at the end the lessor generally does not become the owner Ownership; "the most common method" and becoming popular as lease costs rise; frequently costs least in the long run
Advantages (table, p.254–255) Short-term commitment; high flexibility; no cash up front; (text) can delay the purchase decision; better maintainability since the manufacturer maintains Pre-determined payments for a fixed period; no cash up front; usually better vendor service than under rental; little risk of obsolescence; no capital investment; tax benefits (lease cost as expense, sometimes tax credits) Least cost in the long run; distinct tax advantages for a profit-making firm; a business investment; full control over equipment use
Disadvantages Most expensive option; little control of equipment change; not all vendors favour renting; no tax benefit other than deducting rent as an expense More expensive than purchase; may limit hours or equipment use Risk of obsolescence; permanent commitment; full responsibility for all problems (own maintenance staff or a maintenance contract); immediate and larger cash requirements; locked into the system; loan repayments if financed

Conclusion and summary (pp.256–257). Implementation is "the crowning phase" of the project and can render all earlier effort useless if not done properly; its three aspects are training, converting, and reviewing after implementation.

Code and worked examples

Example 1 — Choose the conversion method (course methods, pp.239–240).

Situation Method Reason
Payroll for a bank; failure would be costly; money available for double running Parallel safest; fall back to old system
Small shop swapping its till program over a weekend; no old system worth keeping Direct abrupt; no parallel activity (risky: no fall-back)
New booking system with new techniques; try in one branch first Pilot working version in one department, feedback, then roll out
Files and trainers can only be ready branch by branch over four months Phase-in staged over weeks to months
Extra cost of running both systems is unaffordable not parallel parallel doubles operating costs

Example 2 — File conversion controls (course p.242: record counts, financial controls, hash totals). Old file of 5 customer records (illustrative data):

Account no. Balance (N)
10234 12,500
10567 3,200
10891 7,800
11023 950
11450 4,100

Control totals taken from the old file: record count = 5; financial control total (sum of balances) = N28,550; hash total (sum of account numbers, meaningless as a quantity but a checksum) = 10234 + 10567 + 10891 + 11023 + 11450 = 54,165.

Case A, a digit transposition: new file has account 10657 in place of 10567. Record count 5 (matches); balances total N28,550 (matches); hash total = 54,255, differs by 90, so the hash total catches the error that the other two missed. Case B, a lost record: new file lacks account 11023 (balance N950): record count 4 (≠5); balances total N27,600 (N950 short); hash total 43,142 (11,023 short). All three controls flag it.

Example 3 — Weighted tender evaluation (course p.248: weights for desirable features; weights and scores are illustrative). Weights: price 40%, performance 25%, spare capacity 15%, vendor support 20% (sum 100%). Scores out of 10:

Bid Price (.40) Performance (.25) Spare capacity (.15) Vendor support (.20) Weighted score
A 7 8 6 9 2.80 + 2.00 + 0.90 + 1.80 = 7.50
B 9 6 7 6 3.60 + 1.50 + 1.05 + 1.20 = 7.35
C 8 8 8 7 3.20 + 2.00 + 1.20 + 1.40 = 7.80

Bid C ranks first. First, as the course says, any bid failing a mandatory requirement is screened out before weighting (p.249).

Example 4 — Renting vs leasing vs purchase (course table pp.254–255; all figures illustrative). A computer costs N2,000,000. Rent N60,000 per month; lease N40,000 per month for 5 years, maintenance included; purchase with a maintenance contract N5,000 per month.

Period Renting Leasing (5-year commitment) Purchase (price + maintenance)
2 years (24 months) 60,000 × 24 = N1,440,000 still committed to N2,400,000 2,000,000 + 5,000 × 24 = N2,120,000
5 years (60 months) 60,000 × 60 = N3,600,000 40,000 × 60 = N2,400,000 2,000,000 + 5,000 × 60 = N2,300,000

Break-even renting vs buying: 60,000 × m = 2,000,000 + 5,000 × m gives m = 36.4 months. Reading: short horizon, rapid growth or no cash → renting (flexibility, most expensive over time); fixed payments, service and little obsolescence risk, no capital → leasing; long horizon → purchase is cheapest ("least cost in the long run"), at the price of obsolescence risk and a large initial outlay. The course adds tax effects not modelled here.

Example 5 — A review plan checklist (course pp.242–244). For a payroll system one year after changeover:

Plan Item reviewed
Administration does payroll still meet the user's objective (pay all staff on time)? are operating costs within budget and savings realised?
Personnel requirements compare operators' performance with objectives; check training performance by test and interview
Hardware compare the terminals' and network's current performance with design specifications
Documentation review are manuals complete, accurate, and conforming to documentation standards?

Common mistakes

  • Parallel vs pilot vs phase-in vs direct. Parallel = both systems at once (safest, doubles cost); direct = abrupt cut-over (no fall-back); pilot = one part of the organisation first (live test); phase-in = staged over weeks to months. A pilot rollout may then use direct or phase-in for the rest.
  • Renting vs leasing vs purchase. Rent 1–12 months, cancel on 30 or 60 days' notice, most expensive; lease 3–7 years, fixed payments including maintenance, lessor usually does not become the owner; purchase is cheapest in the long run but carries obsolescence risk.
  • Special remedies, actual, consequential, liquidated damages, specific performance are different: specific performance fits the case of a system failing acceptance tests because of memory storage (extra memory at nil or nominal cost).
  • MTBF appears twice: in contract guarantees (3.8.7: usable hours, MTBF, maximum repair time) and in equipment characteristics (3.9.3).
  • Control totals. Record counts, financial controls and hash totals (p.242); a hash total sums an identifier field and has no meaning as a quantity.
  • Maintenance names. Corrective, adaptive, perfective (the text misprints adaptive as "adoption(s)"). Perfective is the largest.
  • Vendor-selection criteria are listed "in descending order by importance": economic factors first, then hardware, software, service, reputation.
  • Training: vendor training is off-site and often fee-based; in-house is tailored and cheaper per head but suffers interruptions.

Errata

Page Course says Correct Note
244 "classified as corrective, adoptions or perfective"; "Adoption maintenance means changing the program functions" Corrective, adaptive, perfective Unit 2 Table 2 (p.213) spells it "Adaptive" and defines it as accommodating changes to data, files, hardware and system software. The same p.244 paragraph later says "adaptive".
240–241 "After the system conversion is completed using any one of the methods ..., the conversion plan starts." The conversion plan is prepared before conversion The next page says the plan is made "during the pre-implementation stages, when the conversion is being planned"; Unit 1 (p.190) also describes the conversion plan as preparatory to the changeover.
256 "the user organization takes little to the equipment" takes title to the equipment Wording slip for ownership.
242 "extent to which the system accomplishes stated objectives and actual project costs exceed initial estimates" Garbled; read as: whether objectives are met and how actual costs compare with estimates Not a verified correction of numbers; do not memorise "exceed" as a rule.

Practice questions

  1. (MCQ) The course names three aspects of implementation: training personnel, conversion procedures and: A. system design B. post-implementation review C. benchmarking D. documentation
  2. (MCQ) The conversion method that "doubles operating costs" is: A. direct B. pilot C. parallel D. phase-in
  3. (MCQ) The method in which a working version is installed in one department first, then the rest of the organisation converts all at once or gradually, is: A. parallel B. pilot C. direct D. batch
  4. (MCQ) Which control is useful for catching transposed digits in account numbers during file conversion? A. Record count B. Hash total C. Financial control total of balances D. Site preparation
  5. (MCQ) Which is not one of the four plans in the review plan? A. Administration plan B. Personnel requirements plan C. Hardware plan D. Training brochure plan
  6. (MCQ) Contract guarantees for reliability include minimum hours of usable time per day, maximum time to repair and: A. mean-time-between-failures (MTBF) B. bankruptcy cover C. escalation clauses D. liquidated damages
  7. (MCQ) Renting differs from leasing in that renting: A. is a commitment of three to seven years B. is short-term (1–12 months) and the most expensive option C. gives ownership at the end D. requires full cash up front
  8. (Fill) The three ways of evaluating proprietary software in the course are benchmarking, ______ of other users and ______ of independent research organisations.
  9. (Self-assessment i–iii) (a) List the training methods. (b) List four methods of system conversion. (c) Explain the importance of post-implementation review.
  10. (TMA) (a) What do you know about system maintenance? (b) What are the criteria for vendor selection? (c) List the approaches for software evaluation.
Answersshow
  1. B: p.236.
  2. C: "It doubles operating costs" (p.240).
  3. B: pilot system (p.240).
  4. B: a hash total (sum of account numbers) changes when digits are transposed; record count and balance totals do not (p.242).
  5. D: administration, personnel requirements, hardware, documentation review (pp.242–244).
  6. A: p.247.
  7. B: renting 1–12 months, most expensive (pp.254–255).
  8. experience; report (p.252).
  9. (a) Vendor and in-service training; in-house training (p.238). (b) Parallel systems, direct conversion, pilot system, phase-in (p.239). (c) It measures system performance against pre-defined requirements, shows whether objectives and benefits are met and whether major redesign is needed, and supplies lessons for later projects (p.242, p.243).
  10. (a) The last part of the life cycle, the implementation of the review plan; 60–90% of lifetime software cost; types corrective, adaptive, perfective (perfective largest); average system life 4–6 years; software is corrected rather than worn out; reduce the need by defining requirements accurately, better documentation, effective design methods, better tools, good management (pp.244–245). (b) Economic factors, hardware factors, software factors, service factors, reputation of the manufacturer, in descending order of importance; mandatory requirements screen vendors out (pp.249–250). (c) Benchmarking, experience of other users, report of independent research organisations (p.252).

External resources