Only in research
settings, something that a software development company usually can??™t afford to do,
is there the possibility of having controlled tests over time. As with any case study
where signi?¬?cant bene?¬?ts have been achieved, reporting them in an educational way is
faced with two dif?¬?culties: the danger of sounding triumphalist when providing
Domain-Speci?¬?c Modeling: Enabling Full Code Generation, Steven Kelly and Juha-Pekka Tolvanen
Copyright # 2008 John Wiley & Sons, Inc.
21
evidence from case study participants and the problem of providing enough detail
without revealing trade secrets. Changes in productivity, however, are very visible
since they can be detected in the daily work of developers, especially the change
from using code concepts to using domain concepts, and automatically producing
the code can be drastic: For example, a development task that earlier took days can
be done with DSM in minutes. Consider the difference in using UML and DSM in
the mobile phone application described in Chapter 1. Because the difference is so
clear, development organizations normally have little interest in conducting further
comparisons to calculate return on investment for the creation of a DSM solution.
Later, once the DSM is in place, the organization may ???forget??? the older practices
and there are no longer clear yardsticks available to compare productivity.
Pages:
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59