Eclipse Community Forums: TMF (Xtext) » comments on the proposal

Help

Home

Home » Modeling » TMF (Xtext) » comments on the proposal

Show: Today's Messages :: Show Polls :: Message Navigator

comments on the proposal [message #435]

Thu, 15 November 2007 09:08

Miguel Garcia

Messages: 77
Registered: July 2009

Member

Hi,

I have a few comments on the proposal, as follows:

1) A paragraph in the "Related Issues" section comments on the inability of
Emfatic to adopt any other DSL than the hardcoded Emfatic DSL. That's no
surprise, as Emfatic is not a text-editor-generator, but was itself
generated. If you mention Emfatic at all, please add a link to
http://wiki.eclipse.org/Emfatic where up-to-date information can be found on
the new Emfatic (the one accepted as EMFT component, that improves on the
alphaWorks distribution from a few years ago)

2) In the "Existing Approaches" section, please add a paragraph mentioning
Gymnast Generator, http://wiki.eclipse.org/Gymnast , bundled in the EMFT
Emfatic component. Gymnast Generator was used to generate a large portion of
Emfatic, as a step forward towards combining parser generators with modeling
in a symbiotic way.

3) Finally, I would appreciate if you add the Emfatic team (Miguel Garcia,
Chris Daly) to the list of interested parties for TMF.

Regards,

Miguel

Committer in EMFT Emfatic, MDT OCL Tools, and SOC

--
Miguel Garcia miguel.garcia@tuhh.de
Institute for Software Systems (STS), AB 4-02
Technische Universitaet Hamburg-Harburg
Harburger Schlossstr. 20, 21073 Hamburg Fax: (+49)40-42878-251

Report message to a moderator

Re: comments on the proposal [message #444 is a reply to message #435]

Fri, 16 November 2007 09:00

Eclipse User

Originally posted by: sven.efftinge.de

Hi Miguel,

thanks for the feedback, and sorry for having such an incomplete mention
of Emfatic/Gymnast. Is there some kind of documentation about Gymnast?

We'll update the proposal ASAP.

thanks,
Sven

Miguel Garcia wrote:
> Hi,
>
> I have a few comments on the proposal, as follows:
>
> 1) A paragraph in the "Related Issues" section comments on the inability of
> Emfatic to adopt any other DSL than the hardcoded Emfatic DSL. That's no
> surprise, as Emfatic is not a text-editor-generator, but was itself
> generated. If you mention Emfatic at all, please add a link to
> http://wiki.eclipse.org/Emfatic where up-to-date information can be found on
> the new Emfatic (the one accepted as EMFT component, that improves on the
> alphaWorks distribution from a few years ago)
>
> 2) In the "Existing Approaches" section, please add a paragraph mentioning
> Gymnast Generator, http://wiki.eclipse.org/Gymnast , bundled in the EMFT
> Emfatic component. Gymnast Generator was used to generate a large portion of
> Emfatic, as a step forward towards combining parser generators with modeling
> in a symbiotic way.
>
> 3) Finally, I would appreciate if you add the Emfatic team (Miguel Garcia,
> Chris Daly) to the list of interested parties for TMF.
>
>
> Regards,
>
>
> Miguel
>
>
> Committer in EMFT Emfatic, MDT OCL Tools, and SOC
>
> --
> Miguel Garcia miguel.garcia@tuhh.de
> Institute for Software Systems (STS), AB 4-02
> Technische Universitaet Hamburg-Harburg
> Harburger Schlossstr. 20, 21073 Hamburg Fax: (+49)40-42878-251
>
>

Report message to a moderator

Gymnast patterns and EMF-based syntax trees [message #457 is a reply to message #444]

Fri, 16 November 2007 11:03

Miguel Garcia

Messages: 40
Registered: July 2009

Member

Sven,

The patterns followed in Gymnast-generated text editors are described in
http://www.sts.tu-harburg.de/~mi.garcia/SoC2007/draftreport. pdf
How Gymnast Generator internally works is not documented, however.

That report also describes another generator (IDEalize) that complements
Gymnast, for converting (at runtime) POJO-based Concrete Syntax Trees
into EMF-based ones. With that, those CSTs can be validated or queried
with OCL, or further processed with QVT, for example.

If TMF can do that, too, then we're talking about the same guiding
principles.

Miguel

Report message to a moderator

Re: Gymnast patterns and EMF-based syntax trees [message #470 is a reply to message #457]

Fri, 16 November 2007 16:32

Frédéric Jouault

Messages: 572
Registered: July 2009

Senior Member

Hi Miguel,

> The patterns followed in Gymnast-generated text editors are described in
> http://www.sts.tu-harburg.de/~mi.garcia/SoC2007/draftreport. pdf
> How Gymnast Generator internally works is not documented, however.
>
> That report also describes another generator (IDEalize) that complements
> Gymnast, for converting (at runtime) POJO-based Concrete Syntax Trees
> into EMF-based ones. With that, those CSTs can be validated or queried
> with OCL, or further processed with QVT, for example.
>
> If TMF can do that, too, then we're talking about the same guiding
> principles.

Let me first try to understand the relative positions of Gymnast and TCS.

If I am not mistaken correctly, Gymnast generates a metamodel from a
grammar. It is possible to follow this approach with TMF using
TMF/xText. I will let you (Miguel and Sven) tell me more about the
details ;-).

TMF can also start from a metamodel. TMF/TCS lets you specify the
information necessary to create a grammar that is not already in the
metamodel. Can Gymnast do that as well?
I am surprised that you do not mention TCS in your bibliography.

Also, do you have a list of syntax constructions supported by Gymnast?
For instance, can you have operators and expressions, does the generated
metamodel define a tree or can it also define a graph (e.g., attaching
variable usage to definition/declaration)?

Best regards,

Frédéric Jouault

Report message to a moderator