The CTO of the company saw the change as signi?¬?cant: ???Traditional
programming has largely disappeared and we can build systems up to ?¬?ve times faster
with fewer errors.??? These results were similar to the gains discussed in Chapter 2. In
particular, the automatic generation made the development process easy and safe:
modelers did not need to consider if the correct version of the supporting code was
available during code generation.
The insurance experts could start using the modeling language and related tool
relatively quickly??”after 2 days of training. The main dif?¬?culties lay in learning to
reuse existing product speci?¬?cations or their parts and learning to use the MOFspeci
?¬?c parts visible in the modeling language. Later, analysis of the speci?¬?cations
showed that some of the MOF-speci?¬?c concepts were not needed at all and thus could
be removed: they just led to an unnecessarily complex language.
Particularly impressive to the modelers was their capability to generate working
speci?¬?cations that they could immediately see in the portal. In a similar vein, the sheer
amount of code generated was considerable, 2000??“4000 lines of code per product.
This was almost completely a consequence of the input format used by the underlying
138 INSURANCE PRODUCTS
repository of the web portal.
Pages:
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273