Just de-emphasize the "platform independence" issue, forget about the UML and bin the PSM and any reverse engineering you had in mind.
Industrial experiences of this kind of "MDA" show improvements in productivity and quality with a FACTOR of between 5 and 10 times compared to UML-based MDA implementations.....ehr....hypothetical UML-based implementations of MDA that is :)
Industrial experiences of this kind of "MDA" show improvements in productivity and quality with a FACTOR of between 5 and 10 times compared to UML-based MDA implementations.....ehr....hypothetical UML-based implementations of MDA that is :)
1 comment:
MDA is the OMG's trademark, and means precisely and only the MDA's own Model Driven Architecture. MDA works for some people to some extent, but even tool vendor sponsored lab experiments only show productivity increasing by 35%: http://www.theserverside.com/articles/article.tss?l=SymposiumCoverage#optimalJ
It's also wrong to say there are no UML-based MDA tools. The tool in the study above is just one of several. What's missing compared to the OMG's (updated?) vision is MOF-based MDA tools, which would allow people to create their own Domain-Specific Modeling languages.
When freed from the constraints of the OMG, such languages and their associated domain-specific code generators are what give the 5-10x productivity increases, as you mention (and rightly point to the DSM Forum).
There's a reason why MOF-based MDA tools allowing DSM have not appeared, and will not appear. MOF is fundamentally flawed for designing modeling languages, and in any case only covers a fraction of the things needed in making your own DSM language and generators. The OMG expect you to hand-code the rest - the graphical behaviour of a modeling tool. Once again, as with UML, the OMG are half-solving 10% of a problem, and claiming this is now the standard and all must follow it...
Post a Comment