[Coin-standards] An argument for public domain modeling infrastructure

[Leonardo B. Lopes]

> 
> If I understand you correctly, in our ideal world, we'd have
>    A modeling language LEO, which would use...
>    An interface layer  IRV, which would integrate with...
>    A standard for representing instances

Almost. It is more like we'd have
	An interface layer IRV, to manipulate...
	A *multitude* of modeling *environments*, some of which would look
much like ampl or gams, or like the ilog or ibm libraries, or like
spreadsheet systems, and some of which would look very different. The
modeling environments would generate objects conforming to...
	A standard for representing instances.

>From most to least abstract, that is how I see it. More than one modeling
environment may also collaborate to generate instances. The example I'm
working on is in SP, but I also see domain-specific modeling tools using
the same principals.

> 
> I agree that all 3 would be nice to have.  I think we may just differ on 
> what we think is more valuable in industry.
> 

Apparently we agree even more... I think what you interpreted as IRV is
actually part of the standard for instances. And I don't even think we
disagree on what is most important for industry. The reasons I push for
the instance standard are two: first, it is the component at the lowest
abstraction level, which others will build on; second, it is the easiest.

Cheers,
Leo.

========================================================================
Leonardo B. Lopes                                       leo at iems.nwu.edu 
Ph.D. Student                                              (847)491-8470
IEMS - Northwestern University              http://www.iems.nwu.edu/~leo