Skip to main content


Eclipse Community Forums
Forum Search:

Search      Help    Register    Login    Home
Home » Modeling » EMF "Technology" (Ecore Tools, EMFatic, etc)  » [Announce] EMF Transform proposal
[Announce] EMF Transform proposal [message #132684] Tue, 03 February 2009 13:53 Go to next message
Eclipse UserFriend
Originally posted by: c.krause.cwi.nl

Dear all,

this is an announcement for EMF Transform, a proposed new EMFT
component. EMF Transform is a model transformation framework which is
based on graph transformation techniques.

You can find the project proposal at:

http://tfs.cs.tu-berlin.de/emftrans/proposal

Please post any questions, comments and suggestions to the
eclipse.technology.emft newsgroup with [EMF Transform] prefixed to the
subject. We look forward to your feedback.

Christian
Re: [Announce] EMF Transform proposal [message #132710 is a reply to message #132684] Wed, 04 February 2009 12:38 Go to previous messageGo to next message
Jan Koehnlein is currently offline Jan KoehnleinFriend
Messages: 760
Registered: July 2009
Location: Hamburg
Senior Member
Hi,

I am curious how you think EMF Transform relates to other model
transformation subprojects/components in Eclipse. There are already
several M2M transform languages, e.g. ATL, QVT and Xtend, which can also
be used for in-place transformations. Is your main focus on the
graphical definition of transformations? Do you see overlapping regions
of interest?

Therefore, I feel that the project's name might be a little bit misleading.

OTOH, in-place transformations of Ecore models could also be considered
refactorings, and there's another proposal for that in this list.
news://news.eclipse.org:119/gm9odp$lp3$1@build.eclipse.org
Did you think there are synergies?

Best regards
Jan.


Christian Krause schrieb:
> Dear all,
>
> this is an announcement for EMF Transform, a proposed new EMFT
> component. EMF Transform is a model transformation framework which is
> based on graph transformation techniques.
>
> You can find the project proposal at:
>
> http://tfs.cs.tu-berlin.de/emftrans/proposal
>
> Please post any questions, comments and suggestions to the
> eclipse.technology.emft newsgroup with [EMF Transform] prefixed to the
> subject. We look forward to your feedback.
>
> Christian


---
Get professional support from the Xtext committers at www.typefox.io
Re: [Announce] EMF Transform proposal [message #132752 is a reply to message #132710] Wed, 04 February 2009 14:51 Go to previous messageGo to next message
Eclipse UserFriend
Originally posted by: c.krause.cwi.nl

Hi Jan,

the graphical notation is of course one of the differences to
ATL/QVT/Xtend. More importantly though is the fact that we do not
enforce any concrete notation/syntax (graphical or textual) for our
transformations. Instead, we use an abstract transformation model.

One other difference is the tight coupling with EMF. We focus on EMF
transformations - nothing else (e.g. in ATL they use a different format
for internal representation, called KM3, I believe). All models that we
are dealing with are typed over the Ecore model (including the
transformation model). We can of course also define transformations
for the Ecore model itself (for some examples see the article
http://eceasst.cs.tu-berlin.de/index.php/eceasst/article/vie w/34/22).

I agree with you that refactorings and in-place transformations have a
lot in common. I think the difference that is proposed here is that the
refactorings framework is used to integrate in-place transformations
into editors and in general to provide a framework for the required user
interaction.

Thank you for your comments,
Christian


Jan Kšoehnlein wrote:
> Hi,
>
> I am curious how you think EMF Transform relates to other model
> transformation subprojects/components in Eclipse. There are already
> several M2M transform languages, e.g. ATL, QVT and Xtend, which can also
> be used for in-place transformations. Is your main focus on the
> graphical definition of transformations? Do you see overlapping regions
> of interest?
>
> Therefore, I feel that the project's name might be a little bit misleading.
>
> OTOH, in-place transformations of Ecore models could also be considered
> refactorings, and there's another proposal for that in this list.
> news://news.eclipse.org:119/gm9odp$lp3$1@build.eclipse.org
> Did you think there are synergies?
>
> Best regards
> Jan.
>
>
> Christian Krause schrieb:
>> Dear all,
>>
>> this is an announcement for EMF Transform, a proposed new EMFT
>> component. EMF Transform is a model transformation framework which is
>> based on graph transformation techniques.
>>
>> You can find the project proposal at:
>>
>> http://tfs.cs.tu-berlin.de/emftrans/proposal
>>
>> Please post any questions, comments and suggestions to the
>> eclipse.technology.emft newsgroup with [EMF Transform] prefixed to the
>> subject. We look forward to your feedback.
>>
>> Christian
Re: [Announce] EMF Transform proposal [message #132765 is a reply to message #132710] Wed, 04 February 2009 15:44 Go to previous messageGo to next message
Eclipse UserFriend
Originally posted by: sjurack.mathematik.uni-marburg.de

Hi Jan,

at first, thanks for your interest and feedback. Please find my comments
below.

Jan Kšoehnlein schrieb:
> Hi,
>
> I am curious how you think EMF Transform relates to other model
> transformation subprojects/components in Eclipse. There are already
> several M2M transform languages, e.g. ATL, QVT and Xtend, which can also
> be used for in-place transformations. Is your main focus on the
> graphical definition of transformations? Do you see overlapping regions
> of interest?
Our focus is, on the one hand, to define transformations in a convenient
way, that is using a comprehensible concept of defining
transformations and a supportive graphical editor. Thus, the designer
may not have any knowledge about additional transformation laguages (and
its syntax). On the other hand, using the well-founded theory of graph
transformation in the background enables the exploitation of analysis
features e.g. termination checks and checks in terms of
interdependencies between defined transformations.

As far as I know, ATL does not support in-place transformations, since
it focuses on exogenous model transformations(transformations between
models with different meta-models). We support this as well, in contrast
to ATL we do not need a mapping model as the mapping is done directly
within the rules. However, we anticipate collaborations with other
transformation language projects. E.g. some kind of language conversions
or even language combinations are conceivable in order to utilize
advantages of different languages.

>
> Therefore, I feel that the project's name might be a little bit misleading.
Would you explain this, please? What did you have in mind, when you read
it first time?

>
> OTOH, in-place transformations of Ecore models could also be considered
> refactorings, and there's another proposal for that in this list.
> news://news.eclipse.org:119/gm9odp$lp3$1@build.eclipse.org
> Did you think there are synergies?
Yes, definitely there are synergies. Moreover, as written within the EMF
Refactor proposal, EMF Refactor provides "..an environment of developing
model refactorings based on model transformations defined by EMF
Transform."

Best regards,
Stefan

>
> Best regards
> Jan.
>
>
> Christian Krause schrieb:
>> Dear all,
>>
>> this is an announcement for EMF Transform, a proposed new EMFT
>> component. EMF Transform is a model transformation framework which is
>> based on graph transformation techniques.
>>
>> You can find the project proposal at:
>>
>> http://tfs.cs.tu-berlin.de/emftrans/proposal
>>
>> Please post any questions, comments and suggestions to the
>> eclipse.technology.emft newsgroup with [EMF Transform] prefixed to the
>> subject. We look forward to your feedback.
>>
>> Christian
Re: [Announce] EMF Transform proposal [message #132896 is a reply to message #132765] Mon, 09 February 2009 10:33 Go to previous messageGo to next message
Jan Koehnlein is currently offline Jan KoehnleinFriend
Messages: 760
Registered: July 2009
Location: Hamburg
Senior Member
Stefan Jurack schrieb:
.... snipped
>> Therefore, I feel that the project's name might be a little bit misleading.
>>
> Would you explain this, please? What did you have in mind, when you read
> it first time?
>
Well, that sounds to me like the default technology to transform EMF
models. You should try to reflect the specifics of your transformation
in the project name, like "EMF Graph Transform" or something similar.

Regards
Jan


---
Get professional support from the Xtext committers at www.typefox.io
Re: [Announce] EMF Transform proposal [message #132909 is a reply to message #132896] Mon, 09 February 2009 12:47 Go to previous message
Eclipse UserFriend
Originally posted by: sjurack.mathematik.uni-marburg.de

Hi Jan,

thanks for your suggestion. We will discuss this in our projekt group.

Best regards,
Stefan

Jan Kšoehnlein schrieb:
> Stefan Jurack schrieb:
> ... snipped
>>> Therefore, I feel that the project's name might be a little bit
>>> misleading.
>>>
>> Would you explain this, please? What did you have in mind, when you read
>> it first time?
>>
> Well, that sounds to me like the default technology to transform EMF
> models. You should try to reflect the specifics of your transformation
> in the project name, like "EMF Graph Transform" or something similar.
>
> Regards
> Jan
Re: [Announce] EMF Transform proposal [message #620619 is a reply to message #132684] Wed, 04 February 2009 12:38 Go to previous message
Jan Koehnlein is currently offline Jan KoehnleinFriend
Messages: 760
Registered: July 2009
Location: Hamburg
Senior Member
Hi,

I am curious how you think EMF Transform relates to other model
transformation subprojects/components in Eclipse. There are already
several M2M transform languages, e.g. ATL, QVT and Xtend, which can also
be used for in-place transformations. Is your main focus on the
graphical definition of transformations? Do you see overlapping regions
of interest?

Therefore, I feel that the project's name might be a little bit misleading.

OTOH, in-place transformations of Ecore models could also be considered
refactorings, and there's another proposal for that in this list.
news://news.eclipse.org:119/gm9odp$lp3$1@build.eclipse.org
Did you think there are synergies?

Best regards
Jan.


Christian Krause schrieb:
> Dear all,
>
> this is an announcement for EMF Transform, a proposed new EMFT
> component. EMF Transform is a model transformation framework which is
> based on graph transformation techniques.
>
> You can find the project proposal at:
>
> http://tfs.cs.tu-berlin.de/emftrans/proposal
>
> Please post any questions, comments and suggestions to the
> eclipse.technology.emft newsgroup with [EMF Transform] prefixed to the
> subject. We look forward to your feedback.
>
> Christian


---
Get professional support from the Xtext committers at www.typefox.io
Re: [Announce] EMF Transform proposal [message #620622 is a reply to message #132710] Wed, 04 February 2009 14:51 Go to previous message
Eclipse UserFriend
Originally posted by: c.krause.cwi.nl

Hi Jan,

the graphical notation is of course one of the differences to
ATL/QVT/Xtend. More importantly though is the fact that we do not
enforce any concrete notation/syntax (graphical or textual) for our
transformations. Instead, we use an abstract transformation model.

One other difference is the tight coupling with EMF. We focus on EMF
transformations - nothing else (e.g. in ATL they use a different format
for internal representation, called KM3, I believe). All models that we
are dealing with are typed over the Ecore model (including the
transformation model). We can of course also define transformations
for the Ecore model itself (for some examples see the article
http://eceasst.cs.tu-berlin.de/index.php/eceasst/article/vie w/34/22).

I agree with you that refactorings and in-place transformations have a
lot in common. I think the difference that is proposed here is that the
refactorings framework is used to integrate in-place transformations
into editors and in general to provide a framework for the required user
interaction.

Thank you for your comments,
Christian


Jan Kšoehnlein wrote:
> Hi,
>
> I am curious how you think EMF Transform relates to other model
> transformation subprojects/components in Eclipse. There are already
> several M2M transform languages, e.g. ATL, QVT and Xtend, which can also
> be used for in-place transformations. Is your main focus on the
> graphical definition of transformations? Do you see overlapping regions
> of interest?
>
> Therefore, I feel that the project's name might be a little bit misleading.
>
> OTOH, in-place transformations of Ecore models could also be considered
> refactorings, and there's another proposal for that in this list.
> news://news.eclipse.org:119/gm9odp$lp3$1@build.eclipse.org
> Did you think there are synergies?
>
> Best regards
> Jan.
>
>
> Christian Krause schrieb:
>> Dear all,
>>
>> this is an announcement for EMF Transform, a proposed new EMFT
>> component. EMF Transform is a model transformation framework which is
>> based on graph transformation techniques.
>>
>> You can find the project proposal at:
>>
>> http://tfs.cs.tu-berlin.de/emftrans/proposal
>>
>> Please post any questions, comments and suggestions to the
>> eclipse.technology.emft newsgroup with [EMF Transform] prefixed to the
>> subject. We look forward to your feedback.
>>
>> Christian
Re: [Announce] EMF Transform proposal [message #620623 is a reply to message #132710] Wed, 04 February 2009 15:44 Go to previous message
Stefan Jurack is currently offline Stefan JurackFriend
Messages: 24
Registered: February 2010
Junior Member
Hi Jan,

at first, thanks for your interest and feedback. Please find my comments
below.

Jan Kšoehnlein schrieb:
> Hi,
>
> I am curious how you think EMF Transform relates to other model
> transformation subprojects/components in Eclipse. There are already
> several M2M transform languages, e.g. ATL, QVT and Xtend, which can also
> be used for in-place transformations. Is your main focus on the
> graphical definition of transformations? Do you see overlapping regions
> of interest?
Our focus is, on the one hand, to define transformations in a convenient
way, that is using a comprehensible concept of defining
transformations and a supportive graphical editor. Thus, the designer
may not have any knowledge about additional transformation laguages (and
its syntax). On the other hand, using the well-founded theory of graph
transformation in the background enables the exploitation of analysis
features e.g. termination checks and checks in terms of
interdependencies between defined transformations.

As far as I know, ATL does not support in-place transformations, since
it focuses on exogenous model transformations(transformations between
models with different meta-models). We support this as well, in contrast
to ATL we do not need a mapping model as the mapping is done directly
within the rules. However, we anticipate collaborations with other
transformation language projects. E.g. some kind of language conversions
or even language combinations are conceivable in order to utilize
advantages of different languages.

>
> Therefore, I feel that the project's name might be a little bit misleading.
Would you explain this, please? What did you have in mind, when you read
it first time?

>
> OTOH, in-place transformations of Ecore models could also be considered
> refactorings, and there's another proposal for that in this list.
> news://news.eclipse.org:119/gm9odp$lp3$1@build.eclipse.org
> Did you think there are synergies?
Yes, definitely there are synergies. Moreover, as written within the EMF
Refactor proposal, EMF Refactor provides "..an environment of developing
model refactorings based on model transformations defined by EMF
Transform."

Best regards,
Stefan

>
> Best regards
> Jan.
>
>
> Christian Krause schrieb:
>> Dear all,
>>
>> this is an announcement for EMF Transform, a proposed new EMFT
>> component. EMF Transform is a model transformation framework which is
>> based on graph transformation techniques.
>>
>> You can find the project proposal at:
>>
>> http://tfs.cs.tu-berlin.de/emftrans/proposal
>>
>> Please post any questions, comments and suggestions to the
>> eclipse.technology.emft newsgroup with [EMF Transform] prefixed to the
>> subject. We look forward to your feedback.
>>
>> Christian
Re: [Announce] EMF Transform proposal [message #620633 is a reply to message #132765] Mon, 09 February 2009 10:33 Go to previous message
Jan Koehnlein is currently offline Jan KoehnleinFriend
Messages: 760
Registered: July 2009
Location: Hamburg
Senior Member
Stefan Jurack schrieb:
.... snipped
>> Therefore, I feel that the project's name might be a little bit misleading.
>>
> Would you explain this, please? What did you have in mind, when you read
> it first time?
>
Well, that sounds to me like the default technology to transform EMF
models. You should try to reflect the specifics of your transformation
in the project name, like "EMF Graph Transform" or something similar.

Regards
Jan


---
Get professional support from the Xtext committers at www.typefox.io
Re: [Announce] EMF Transform proposal [message #620634 is a reply to message #132896] Mon, 09 February 2009 12:47 Go to previous message
Stefan Jurack is currently offline Stefan JurackFriend
Messages: 24
Registered: February 2010
Junior Member
Hi Jan,

thanks for your suggestion. We will discuss this in our projekt group.

Best regards,
Stefan

Jan Kšoehnlein schrieb:
> Stefan Jurack schrieb:
> ... snipped
>>> Therefore, I feel that the project's name might be a little bit
>>> misleading.
>>>
>> Would you explain this, please? What did you have in mind, when you read
>> it first time?
>>
> Well, that sounds to me like the default technology to transform EMF
> models. You should try to reflect the specifics of your transformation
> in the project name, like "EMF Graph Transform" or something similar.
>
> Regards
> Jan
Previous Topic:[Teneo] Problem with saving after deleting object referencing StringToXXXMap
Next Topic:[EMF Compare] Troubles with multiple merge operations in the Eclipse session
Goto Forum:
  


Current Time: Thu Mar 28 15:26:32 GMT 2024

Powered by FUDForum. Page generated in 0.03128 seconds
.:: Contact :: Home ::.

Powered by: FUDforum 3.0.2.
Copyright ©2001-2010 FUDforum Bulletin Board Software

Back to the top