9.7 MAIN RESULTS
Comparing the Watch models to the tangle of code that is normally found when
similar embedded systems are hand coded, it is clear that the DSM solution helped
build systems better here than is normal in the industry. However, looking at the
generated applications, it is equally clear that the code produced is shorter and in some
ways simpler than would normally be written by hand. How much then is the
improvement due to DSM actually only due to the greater attention spent on
developing a good framework?
To gain some insight into this, we devised an experiment. Subjects extended the
Stopwatch application to add lap time functionality, ?¬?rst by DSM with code
generation, and again manually by editing the original Stopwatch Java code in the
same architecture. While the sample size was too small to be statistically signi?¬?cant,
the results in Table 9.1 may still be interesting.
The necessary changes to add a lap time function were roughly eight lines of Java
code or eight objects in the model. For a senior developer, the productivity for
222 DIGITAL WRISTWATCH
modeling was over ?¬?ve times that for coding. In the case of a junior developer, the
difference was only four times, but for the combination of both developers the
productivity for modeling was ?¬?ve times that for coding.
Pages:
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405