21.01.2014 Views

A Rationale-based Model for Architecture Design Reasoning

A Rationale-based Model for Architecture Design Reasoning

A Rationale-based Model for Architecture Design Reasoning

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

5.3. Survey findings<br />

previous design decisions (Table 5.13), it can be argued that the availability of design<br />

rationale can improve this process. <strong>Design</strong> rationale is considered an important input to<br />

impact analysis because if the knowledge about the factors that may have influenced the<br />

original design decisions is available, designers do not have to spend a large amount of time<br />

to understand the existing design and the potential ramifications of any changes [128].<br />

Table 5.14: Frequency of Per<strong>for</strong>ming Impact Analysis<br />

Never<br />

Always<br />

1 2 3 4 5<br />

No of Respondents 0 4 13 29 19<br />

Percentages 0.0 6.2 20.0 44.6 29.2<br />

We were also interested in determining the importance of the tasks per<strong>for</strong>med in impacts<br />

analysis, some of which require design rationale support. The importance of such<br />

tasks during impact analysis can be considered as an indicator of the need or usefulness<br />

of design rationale to improve this activity and subsequently the maintenance process. To<br />

investigate this issue, there was a multiple-items question on the importance of various<br />

tasks. Table 5.15 shows what the respondents think of the importance of the factors. In<br />

carrying out impact analysis, most respondents trace requirements. They analyse the design<br />

and design rationale <strong>based</strong> on the available design specifications. In the analysis, they<br />

are concerned about the feasibility of the implementation, its costs and risk. They are<br />

also concerned that the new design is consistent with the constraints and the assumptions<br />

of the existing system.<br />

Table 5.15: Importance of Each Impact Analysis Task<br />

Not<br />

Very<br />

Important<br />

Important<br />

1 (%) 2 (%) 3 (%) 4 (%) 5 (%)<br />

(a) Analyse and trace requirements 1.5 0.0 10.3 39.7 48.5<br />

(b) Analyse specifications of previous design 1.5 10.3 19.1 35.3 33.8<br />

(c) Analyse design rationale of previous design 1.5 11.9 32.8 32.8 21.0<br />

(d) Analyse implementation feasibility 1.5 1.5 8.8 41.2 47.0<br />

(e) Analyse violation of constraints and assumptions 1.5 8.8 25.0 41.2 23.5<br />

(f) Analyse scenarios 1.5 7.6 16.7 36.4 37.8<br />

(g) Analyse cost of implementation 0.0 10.3 13.2 30.9 45.6<br />

(h) Analyse risk of implementation 0.0 4.4 5.9 32.4 57.3<br />

Given the results, it is obvious that architects and designers often have the need to<br />

refer to design documentation <strong>for</strong> impact analysis. Some of in<strong>for</strong>mation that they use<br />

require design rationale support such as understanding design assumptions, constraints,<br />

cost and risk. However, this in<strong>for</strong>mation are not necessarily organised in a way that can<br />

be easily used <strong>for</strong> impact analysis. Additional factors that have been discovered in this<br />

survey can contribute to the difficulties in impact analysis: (a) architects can <strong>for</strong>get the<br />

reasons behind the design; (b) the person doing the maintenance may not be the original<br />

designers.<br />

72

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!