[skoll] GME

[Don Hinton]

Hi Bala:

On Mon, 8 Mar 2004, Balachandran Natarajan wrote:

> Don-
> 
> 	I was just discussing with Arvind about this. Here are some
> questions  that we thought are relevant in this context
> 
> 1. Are users of GME from ESCHER going to develop GME by say adding 
>    extra modules?

Sure, why not?

> 
> 2. Are users of GME from ESCHER going to develop metamodels using GME?

Again, sure, why not?

> 
> 3. Will the users of ESCHER also download some metamodels that are part
>    of the repository and use it to develop applications?

And again, why not?

> 
> 4. Repeat question 3 for interpreters.

Sure, and probably for lots of stuff we haven't even thought of.

> 
> Looks like the test cases for the above three would be totally 
> orthogonal to each other. You may want to pick and decide based on what
> the ESHCER is supposed to be housing. The problem gets complicated since
> GME AFAWK is not scriptable.  Just our $0.02.

Okay, and how does that effect the use of Skoll?

We provide a framework, the development team provides the tests.  If we
need/want to test interoperability between disparate tools, we'll have to
add tests.  I don't see this as any different from anything else.

But perhaps I've misunderstood.  Why do you think Skoll wouldn't be a good 
tool to use to automate the GME nightly builds (that currently don't 
exist)?

take care...
don

> 
> Thanks
> Bala
> 
> > Here's a link to GME.  If you could download it and take a look, that
> > would be great.  Also, we have access to the CVS repository on ESCHER
> > (thought I may need to add you to the group).  Anyway, there are several 
> > projects that need to be built, not just GME itself, but there are not 
> > compile-time options to set, just a release and a debug target.
> > 
> > http://www.isis.vanderbilt.edu/Projects/gme/
> > 
> > It's VC6, so a windows skoll environment that uses VC6 and can invoke a 
> > suite of projects as a group would be great.
> > 
> > As for testing, they don't really have any, but if we could come up with 
> > an easy way for them to integrate them, we could test GME under 3 
> > scenerios:
> > 
> > 1) the various GUI's
> > 2) the COM APIs
> > 3) interactions between the tools, a la, tool chains (a very important 
> > concept in ESCHER)
> > 
> > thanks...
> > don
> > 
> > _______________________________________________
> > skoll mailing list
> > skoll at list.isis.vanderbilt.edu
> > http://list.isis.vanderbilt.edu/mailman/listinfo/skoll
> > 
>