The Model Driven Software Network
Raise your level of abstraction
How to make MDSD more attractive to the "mere developer mortal"?
Taking on a good suggestion from Jon Hurwitz, I would like to introduce the above question.
In a recent Metacase webinar about DSM, I asked Juha-Pekka Tolvanen who was MetaEdit+'s biggest competitor. His answer was: "developers". (By the way, you can still attend the Metacase webinar at http://www.metacase.com/webinar/WebinarFeb2009.html).
So one of DSM's thinking leaders recognizes that convincing developers of DSM/MDSD's advantages is the greatest hurdle.
Therefore, MDSD evangelism is required! How do you think this can be achieved? Seminars, Webinars, a great Powerpoint presentation model, hands-on videos, case studies?
In a recent Metacase webinar about DSM, I asked Juha-Pekka Tolvanen who was MetaEdit+'s biggest competitor. His answer was: "developers". (By the way, you can still attend the Metacase webinar at http://www.metacase.com/webinar/WebinarFeb2009.html).
So one of DSM's thinking leaders recognizes that convincing developers of DSM/MDSD's advantages is the greatest hurdle.
Therefore, MDSD evangelism is required! How do you think this can be achieved? Seminars, Webinars, a great Powerpoint presentation model, hands-on videos, case studies?
Replies to This Discussion
-
Permalink Reply by Ed Merks on
-
I think part of the solution is to tackle the problem head on. Identify all the stupid reasons people have for why they think modeling is stupid http://ed-merks.blogspot.com/2008/09/unbearable-stupidity-of-modeli... and then address each of those specific issues. I.e., demonstrate to people where their thinking has gone astray.
The complexity argument is one that I'm quite weary of hearing. Of course I see it repeated in this very note thread too. Go figure. Did we stop to consider that when those who advocate MDD go around pointing fingers at various MDD technologies and argue that the thing over there is too complicated, we're lending weight to the whole complexity argument? I'm sure you have something much simpler. But consider that things always start out much simpler, they don't stay that way. See Java as a case in point: how come enums were too complex to include initially and yet there they are in 5.0? Simple things are more likely not to be very complete and while simple to learn, will tend to feel restrictive and limiting by the time a really complicated task needs to be addressed. Complex things, while complete, will be more challenging to learn, but will be more likely to pan out into a complete solution for the complex problems that mere developer mortals are faced with every day. Where's the Goldilocks point?
EMF continues to grow in popularity for an important reason. It has direct pragmatic value to the mere developer mortal. They don't need to drink the MDD koolaide first. When I asked the platform UI guys what convinced them EMF was a good thing for e4, they said they measured the performance of a reflective eGet verses a simple HashMap get and found it to be twice as fast. Of course the footprint is dramatically smaller and somehow type safety comes apparently for free. Pragmatic value wins the day. Of course it helps a great deal that EMF is also very useful for building the MDD technology itself, i.e., to build other meta models, like UML, and to build DSLs, like Xtext does.
- ‹ Previous
- 1
- 2
- Next ›
About
Badge
© 2010 Created by Mark Dalgarno on Ning. Create a Ning Network!
