Login | Register
My pages Projects Community openCollabNet

Discussions > dev > RE: [gef-dev] model driven GEF

gef
Discussion topic

Hide all messages in topic

All messages in topic

Re: [gef-dev] model driven GEF

Reply

Author Alex Bagehot <alex at bagehot dot net>
Full name Alex Bagehot <alex at bagehot dot net>
Date 2003-05-20 06:45:51 PDT
Message I seem to have made a mistake in that it isn't xmi-di but more uml-di.

I have copied this email to the address given in the proposal. I hope
that someone at Gentleware would find this discussion interesting. I
guess most of the points here would have been discussed as part of
writing the proposal itself so it may seem trivial.

The downsides:

 From reading the proposal, all the diagram details are saved as
'properties' in the uml-di model. That confirms your fear that it is
another standard.

But then the immediate question that pops into my head is what is JGX
(maybe its not fair(/like-with-like) to compare JGX with uml-di?)?
You'll have to forgive me I haven't read up on what is exactly saved in
the JGraphPad file format, but it seems to save the graphmodel and
presentation info (below).

"The first section is a
textual representation of the graphmodel and its group-structure. The
second and third
sections are referenced by the cells in the first part, they hold the
userObjects and
attributes respectively."
http://sourceforge.n​et/mailarchive/forum​.php?thread_id=23612​91&forum_id=7889​

[ArgoUml saves pgml. Each group is assigned a unique id which is also
stored in the model, thus providing the link for save/load. the other
major difference is that ArgoUml doesn't use ports. there are other
problems.]

(i'm not commenting on whether any approach is good or bad) Like you I
am trying to keep an open mind.

I am trying to brainstorm ideas, so I expect to get things wrong. Put it
another way, it is cheaper to make a mess of proposals, suggestions,
ideas now than later when designs are made and code committed.

regards,
Alex

Michael Lawley wrote:
> Alex Bagehot wrote:
>
>
>>what do you think of the xmi diagram interchange proposal from gentleware?
>
>
>>is it better for frameworks like GEF and JGraph to provide support for
>>that rather than focussing on mof or emf?
>
>
>>It would seem to me to be yes.
>
>
> I don't have a strong feeling on this one, but I would tend toward an
> approach that embedded SVG with the model using namespaces.
>
> The advantage of using a diagram model such as Gentleware et al propose
> is that you can leverage MOF/EMF to get both XMI for your domain model
> and for the diagram model (EMF then allows these parts to be stored in
> the one document or split between documents (Resources)).
>
>
> The downside is that a real diagram has a lot more than just boxes and
> lines, nodes and edges -- there's text (with fonts and styling),
> connectedness of lines and possibly shapes, containment (grouping),
> shading, etc etc. Do we really want yet another standard defining all
> these things?
>
> |v| "analog - the new digital"
>
> --
> Michael Lawley, http://purl.org/NET/lawley
> Scientician.
>
>
>
>
>
> --------------------​--------------------​--------------------​---------
> To unsubscribe, e-mail: dev-unsubscribe at gef dot tigris dot org
> For additional commands, e-mail: dev-help at gef dot tigris dot org
>
>



--------------------​--------------------​--------------------​---------
To unsubscribe, e-mail: dev-unsubscribe at gef dot tigris dot org
For additional commands, e-mail: dev-help at gef dot tigris dot org

Re: [gef-dev] model driven GEF

Reply

Author Michael Lawley <lawley at dstc dot edu dot au>
Full name Michael Lawley <lawley at dstc dot edu dot au>
Date 2003-05-19 19:02:49 PDT
Message Alex Bagehot wrote:

> what do you think of the xmi diagram interchange proposal from gentleware?

> is it better for frameworks like GEF and JGraph to provide support for
> that rather than focussing on mof or emf?

> It would seem to me to be yes.

I don't have a strong feeling on this one, but I would tend toward an
approach that embedded SVG with the model using namespaces.

The advantage of using a diagram model such as Gentleware et al propose
is that you can leverage MOF/EMF to get both XMI for your domain model
and for the diagram model (EMF then allows these parts to be stored in
the one document or split between documents (Resources)).


The downside is that a real diagram has a lot more than just boxes and
lines, nodes and edges -- there's text (with fonts and styling),
connectedness of lines and possibly shapes, containment (grouping),
shading, etc etc. Do we really want yet another standard defining all
these things?

|v| "analog - the new digital"

--
    Michael Lawley, http://purl.org/NET/lawley
    Scientician.





--------------------​--------------------​--------------------​---------
To unsubscribe, e-mail: dev-unsubscribe at gef dot tigris dot org
For additional commands, e-mail: dev-help at gef dot tigris dot org

RE: [gef-dev] model driven GEF

Reply

Author hallvard
Full name Hallvard Trætteberg
Date 2003-05-19 06:30:01 PDT
Message Alex,

> what do you think of the xmi diagram interchange proposal from
gentleware?

It's OK as an exchange format, but not as a native format.

> is it better for frameworks like GEF and JGraph to provide
> support for that rather than focussing on mof or emf?

It makes less sense to use MOF or EMF for presentation aspects, since
it's less standardized and normally requires greater speed. So I would
guess a file should consist of one model part, with XMI of the model,
one native presentation/layout-oriented part, and one diagram exchange
part (or may be exported if desired?).

Hallvard


--------------------​--------------------​--------------------​---------
To unsubscribe, e-mail: dev-unsubscribe at gef dot tigris dot org
For additional commands, e-mail: dev-help at gef dot tigris dot org

Re: [gef-dev] model driven GEF

Reply

Author Alex Bagehot <alex at bagehot dot net>
Full name Alex Bagehot <alex at bagehot dot net>
Date 2003-05-19 04:32:21 PDT
Message Hi,

what do you think of the xmi diagram interchange proposal from gentleware?

is it better for frameworks like GEF and JGraph to provide support for
that rather than focussing on mof or emf?

It would seem to me to be yes.

I haven't seen it discussed anywhere.

regards,
Alex

ps. your webpage URL link no longer works.

Michael Lawley wrote:
  > Alex Bagehot wrote:
  >
  >
  >>I recently read this article:
  >
  >
  >>http://www.research.​ibm.com/AEM/eum.html​
  >
  >
  >>It basically says that a modelling tool's graphics framework should have
  >>MOF at its core.
  >
  >
  > Better still, it should have MOF2 (eg EMF as a close approximation) at
  > its core.
  >
  > |v| "analog - the new digital"
  >
  > --
  > Michael Lawley, http://purl.org/NET/lawley
  > Scientician.
  >
  >
  >
  >
  >
  > --------------------​--------------------​--------------------​---------
  > To unsubscribe, e-mail: dev-unsubscribe at gef dot tigris dot org
  > For additional commands, e-mail: dev-help at gef dot tigris dot org
  >
  >





--------------------​--------------------​--------------------​---------
To unsubscribe, e-mail: dev-unsubscribe at gef dot tigris dot org
For additional commands, e-mail: dev-help at gef dot tigris dot org

Re: [gef-dev] model driven GEF

Reply

Author Michael Lawley <lawley at dstc dot edu dot au>
Full name Michael Lawley <lawley at dstc dot edu dot au>
Date 2003-05-18 15:49:34 PDT
Message Alex Bagehot wrote:

> I recently read this article:

> http://www.research.​ibm.com/AEM/eum.html​

> It basically says that a modelling tool's graphics framework should have
> MOF at its core.

Better still, it should have MOF2 (eg EMF as a close approximation) at
its core.

|v| "analog - the new digital"

--
    Michael Lawley, http://purl.org/NET/lawley
    Scientician.





--------------------​--------------------​--------------------​---------
To unsubscribe, e-mail: dev-unsubscribe at gef dot tigris dot org
For additional commands, e-mail: dev-help at gef dot tigris dot org

[gef-dev] model driven GEF

Reply

Author Alex Bagehot <alex at bagehot dot net>
Full name Alex Bagehot <alex at bagehot dot net>
Date 2003-05-16 12:19:43 PDT
Message Hi,

I recently read this article:

http://www.research.​ibm.com/AEM/eum.html​

It basically says that a modelling tool's graphics framework should have
MOF at its core.

Will this influence the direction of GEF?

cheers,
alex



--------------------​--------------------​--------------------​---------
To unsubscribe, e-mail: dev-unsubscribe at gef dot tigris dot org
For additional commands, e-mail: dev-help at gef dot tigris dot org
Messages per page: