This is different from using PGML in two significant ways. Firstly, and
this was the context in which this was written, it has a XMIWriter and
XMIParser generated from a model. Secondly, it forces us to define the
distinction between the two parts of the diagram code that I think
currently are entangled with each-other. The model part (i.e. the code
containing the actual data of the diagrams, their consistency and
storage). The view part (i.e. the code that actually renders the
diagrams, mostly GEF but also all UML-specific functions).
I agree that this is not something we should rush into. Especially since
you now think that there will be a solution available for us within two
years. I did interpret some of your other mails less optimistically than
that.
So, we will be standing here at the roadblock patching our PGML
generator/parser until the road clears...
/Linus
> -----Original Message-----
> From: Tom Morris [mailto:[EMAIL PROTECTED]
> Sent: den 23 september 2006 19:30
> To: [email protected]
> Subject: RE: [argouml-dev] Keeping zargo file compatible with previous
> versions of ArgoUML
>
> Linus suggested:
>
> > Make and maintain a Diagram Interchange model for MOF 1.4 (or
> > as close to the real thing as possible). Load it into the
> > current version of MDR, the one we have installed. XMI and
> > load and save are generated from MDR with old versions.
> >
> > This would solve the file compatibility and saving and
> > loading problems but we still haven't made the transition to
> > UML 2.0 and we're still non-standard.
>
> I don't understand how this is any different from wedging PGML into
XMI or
> any other ArgoUML proprietary approach. Swapping one proprietary
approach
> for another has no advantage that I can see. It causes us to waste
effort
> on an evolutionary dead end throw away piece of code, It doesn't
improve
> interchange with other tools. It replaces a set of known problems
with a
> new set of unknown problems.
>
> Our current proprietary approach, for all of its problems, is at least
a
> known quantity. The only way I'd support replacing it with a
different
> proprietary approach of our own crafting would be if it was clear that
it
> was going to be at least two years before we could get to a standard
UML
> 2.x
> Diagram Interchange.
>
> We've made due with PGML for 5+ years. Why the sudden rush to replace
it
> with something that will have a lifetime measured in months?
>
> Tom
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [EMAIL PROTECTED]
> For additional commands, e-mail: [EMAIL PROTECTED]
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
