Software Design Methodology

Software Design Methodology 255

y Scenario 7. To operate interactively on the input: to allow the user to insert

and delete lines from the original list of lines.

y Scenario 8. To operate interactively on the output: to allow the user to insert

and delete lines from the circular shifted lists.

Whether a scenario is to be used in the SAAM analysis depends on whether

the scenario is considered as important.

Figure 10.3 Quality concerns of KWIC

10.2.2 Description of candidate architecture

The candidate architecture or architectures should be described in an architectural

notation that is well understood by the parties involved in the analysis. These

architectural descriptions must indicate the system’s computation and data

components as well as all relevant connections. Accompanying this description of

the architecture is a description of how the system behaves over time, or a more

dynamic representation of the architecture. This may take the form of a natural

language specification of the overall behaviour or some other more formal and

structured specification.

The KWIC Index System: Quality Concerns

The software that calculates keywords in context should also be

efficient in both space and time. Its components should serve as

reusable entities. It should be able to support the following changes in

the future.

1. Changes in the processing algorithm: For example, line shifting

can be performed on each line as it is read from the input device,

on all the lines after they are read, or on demand when the

alphabetisation requires a new set of shifted lines.

2. Changes in data representation: For example, lines, words, and

characters can be stored in various ways. Similarly, circular shifts

can be stored explicitly or implicitly (as pairs of index and

offset).

3. Enhancement to system function: For example, modify the system to

eliminate circular shifts that start with certain noise words (such as

a, an, and, etc.), change the systems to be interactive, and allow the

user to delete lines from the original lists (or from the circular

shifted lists).

256 Chapter 9. Analysis and Evaluation of Modifiability: The SAAM Method

For the KWIC index system, four architectural designs have been proposed in

the literature and investigated by Mary Shaw and David Garlan in their case

studies. Here, we use two of them, the shared data architecture and the abstract data

type architecture, as the examples to illustrate the subsequent analysis activities.

The architectural design of the KWIC index system in the

main-program-subroutine with shared data style is depicted in Figure 10.4. It

decomposes the system according to the four basic functions performed by the

system: input, shift, alphabetise and output. These computational components are

coordinated as subroutines by a main program that sequences through them in turn.

Data is communicated between the components through shared storage (‘core

storage’). Communication between the computational components and the shared

data is an unconstrained read-write protocol. This is possible because the

coordinating program guarantees sequential access to the data.

Figure 10.4 Design of KWIC in shared-memory architecture

This solution allows data to be represented efficiently, since computations can

share the same storage. The original input of the sequence of lines of characters is

stored in the shared data storage Characters. A circular shift of a line is represented

by a pair of indexes <line, character> and stored in the data storage Index. For

each pair of indexes <line, character>, the line is an integer that gives the position

of the first character of the line in the data storage Characters. The character is an

integer that gives the position of the first character of the shifted line. This

representation avoids the duplication of all the characters in the original input.

Alphabetised Index is an array of integers. Each integer points to a circular shifted

line in the Index. The solution also has a certain intuitive appeal, since distinct

computational aspects are isolated in different modules.

However, as Parnas argued, it has a number of serious drawbacks in terms of

its ability to handle changes. In particular, a change in the data storage format will

KWIC Index System: Architectural Design

in the Main-Program-Subroutine with Shared Data Style

Master Control

Characters

Index Alphabetised Index

Input Circular Shift Alphabetise Output