Comments

# Cyril said on February 7, 2008 12:12 AM:

Mmm.

Just thought you might be interested in what I stumbled upon yesterday. I was surfing around to inform myself even more about MDD, MDA, DSM, DSL, etc-related real world experiences involving code generation "in the large" and I found this amongst other things :

http://www.st.ewi.tudelft.nl/~basgraaf/publications/ATEM2006.pdf

At first, I found it interesting as a kind of an anti-use case of the DSM/DSL approach, that is, to put it in a "versus the good old UML/MOF side of the modeling world" fashion, to have solid rationale about why / when / how to do so -- i.e, not to use DSM/DSL and to stick to the UML realm instead.

Please note I don't deny the fact those folks very likely know what they are talking about as far as UML/MOF/XMI etc are concerned, but I have the feeling we should be much more careful regarding their understanding of the DSM/DSL's *actual* capabilities when they're used properly enough.

After reading at least two times the paper I ended up with the feeling that they confused issues of the very design and implementation of that DSM/DSL-based INVINCIBLE Software Factory with what really pertains to Domain-Specific Languages' features and properties.

My own analysis of their report about the state in which they found that INVINCIBLE Factory is the factory's design was certainly too much tightly- coupled with the targeted runtime architecture (which, also, was sadly a C/S one, no multi-tier) and it would have probably been enough to reengineer that factory "a bit", without having to drop the entire DSM/DSL-based approach so fast.

Instead, they suggested, and eventually decided to virtually blame the latter all, to drop it and to switch to an MDA/UML-based one which I think hasn't brought a big benefit in terms of simplicty, or in other aspects. Just read about what they had to go through while migrating the legacy modeling tools to the new ones...

Well, at least they got the job done, but looks like to me they used a bazooka to fight with Fortis' MDD issues where they could have done it using only a small hand gun.

That's the kind of paper which has many qualities and rich information overall, but which also holds a number of highly arguable (if not false) statements, if you don't look close enough.

I can provide you elsewhere samples of statements I find suspect (or just plain wrong, imho) elsewhere if you ask. Don't want to make this any longer here.

Cheers.

# stuart kent's blog said on February 7, 2008 8:07 AM:

Gareth is obviously settled in the US now, because he's starting to become a prolific blog poster again.

# Noticias externas said on February 7, 2008 8:31 AM:

Gareth is obviously settled in the US now, because he's starting to become a prolific blog poster

# Cyril said on February 9, 2008 9:06 AM:

So, Steve has kindly shared his thoughts with me about this. :)

I now have some more significant input on these issues, should I have to discuss them with those in my company who think it isn't even worth taking the trouble/time to evaluate the usefulness of the DSM/DSL approach vs. the MDA/UML-based standards and tools, albeit I performed something concrete and useful for them with my little AppBuilder, built thanks to your DSL Tools.

New Comments to this post are disabled

Search

Go

This Blog

• Home
• Email

Disclaimer

The information in this weblog is provided "AS IS" with no warranties, and confers no rights. This weblog does not represent the thoughts, intentions, plans or strategies of my employer. It is solely my opinion. Inappropriate comments will be deleted at the authors discretion.
All code samples are provided "AS IS" without warranty of any kind, either express or implied, including but not limited to the implied warranties of merchantability and/or fitness for a particular purpose.