Eclipse Community Forums
Forum Search:

Search      Help    Register    Login    Home
Home » Modeling » GMT (Generative Modeling Technologies) » Re: km3 -> ecore - bad practice in metamodeling?
Re: km3 -> ecore - bad practice in metamodeling? [message #381751] Thu, 20 March 2008 18:53 Go to next message
Ed Merks is currently offline Ed MerksFriend
Messages: 26287
Registered: July 2009
Senior Member
Matti,

I'm not sure what to say. This issue came up recently on the GMT
newgroup and it's something that KM3 folks need to consider. Something
like annotations to specify things not in KM3 but needed in Ecore would
seem to make sense.


Matti Husu wrote:
> Hi, please tell me what do you think and is there a way to avoid the
> problems below.
>
> First you specify metamodel with KM3 there has to be Primitive types
> package if classes have attributes (String, Boolean...).
>
> Then you inject it to a .ecore file. Everything is cool. But if you
> want to make an editor (EMF/GEF or GMF) you need .genmodel.
> This can't be created at least on Eclipse 3.2.1/EMF 2.2.1
>
> In Eclipse 3.3.1.1 you can but you have to specify nsUri etc for the
> PrimitiveTypes-package and corresponding java classes for Integer etc.
>
> I think GMF-ecore diagram editor is more handy.
>
> -Matti
>
Re: km3 -> ecore - bad practice in metamodeling? [message #381752 is a reply to message #381751] Thu, 20 March 2008 19:06 Go to previous messageGo to next message
Dimitrios Kolovos is currently offline Dimitrios KolovosFriend
Messages: 1776
Registered: July 2009
Senior Member
Hi Matti,

Have you tried Emfatic? It is quite similar to KM3 but EMF-specific.

Cheers,
Dimitrios


Ed Merks wrote:
> Matti,
>
> I'm not sure what to say. This issue came up recently on the GMT
> newgroup and it's something that KM3 folks need to consider. Something
> like annotations to specify things not in KM3 but needed in Ecore would
> seem to make sense.
>
>
> Matti Husu wrote:
>> Hi, please tell me what do you think and is there a way to avoid the
>> problems below.
>>
>> First you specify metamodel with KM3 there has to be Primitive types
>> package if classes have attributes (String, Boolean...).
>>
>> Then you inject it to a .ecore file. Everything is cool. But if you
>> want to make an editor (EMF/GEF or GMF) you need .genmodel.
>> This can't be created at least on Eclipse 3.2.1/EMF 2.2.1
>>
>> In Eclipse 3.3.1.1 you can but you have to specify nsUri etc for the
>> PrimitiveTypes-package and corresponding java classes for Integer etc.
>>
>> I think GMF-ecore diagram editor is more handy.
>>
>> -Matti
>>
Re: km3 -> ecore - bad practice in metamodeling? [message #381753 is a reply to message #381752] Thu, 20 March 2008 21:45 Go to previous messageGo to next message
Frédéric Jouault is currently offline Frédéric JouaultFriend
Messages: 572
Registered: July 2009
Senior Member
Hello,

> Have you tried Emfatic? It is quite similar to KM3 but EMF-specific.

.... and therefore Java-specific.

>> I'm not sure what to say. This issue came up recently on the GMT
>> newgroup and it's something that KM3 folks need to consider.
>> Something like annotations to specify things not in KM3 but needed in
>> Ecore would seem to make sense.

Indeed, this is how it works. Here is for instance the ATL annotation
model that complements the ATL KM3 metamodel:
http://dev.eclipse.org/viewcvs/index.cgi/org.eclipse.m2m/org .eclipse.m2m.atl/dsls/ATL/Metamodel/ATL.ann?root=Modeling_Pr oject&view=markup


Regards,

Frédéric Jouault
Re: km3 -> ecore - bad practice in metamodeling? [message #381754 is a reply to message #381753] Thu, 20 March 2008 22:14 Go to previous messageGo to next message
Dimitrios Kolovos is currently offline Dimitrios KolovosFriend
Messages: 1776
Registered: July 2009
Senior Member
Hi Frédéric,

Frédéric Jouault wrote:
> Hello,
>
>> Have you tried Emfatic? It is quite similar to KM3 but EMF-specific.
>
> ... and therefore Java-specific.

Indeed. However, I felt that practicality - as opposed to
purity/technology independence - was the main concern in the original
post, and Emfatic provides built-in support for the required features.
>
>>> I'm not sure what to say. This issue came up recently on the GMT
>>> newgroup and it's something that KM3 folks need to consider.
>>> Something like annotations to specify things not in KM3 but needed in
>>> Ecore would seem to make sense.
>
> Indeed, this is how it works. Here is for instance the ATL annotation
> model that complements the ATL KM3 metamodel:
> http://dev.eclipse.org/viewcvs/index.cgi/org.eclipse.m2m/org .eclipse.m2m.atl/dsls/ATL/Metamodel/ATL.ann?root=Modeling_Pr oject&view=markup
>
>
>
> Regards,
>
> Frédéric Jouault

Greetings from York,
Dimitrios
Re: km3 -> ecore - bad practice in metamodeling? [message #381755 is a reply to message #381752] Fri, 21 March 2008 09:01 Go to previous messageGo to next message
Eclipse UserFriend
Originally posted by: mthusu.cs.helsinki.fi

Hi, not yet. I read about it and it's supposed to be good wih EMF. I will
try it. Right now I'm going to work with my thesis with the tools I have
chosen. I think I chose "KM3 over Emfatic" because KM3 is closely related
to ATL/AMW. For practical metamodeling Emfatic is maybe better. One way is
also to use GEMS. You get editor and all in just minutes. But... the
metamodel gets "polluted" with some mysterious stuff (Memento etc) and
diagram location stuff. I really don't understand why they haven't
separated this stuff from actual metamodel.

-Matti
Re: km3 -> ecore - bad practice in metamodeling? [message #381756 is a reply to message #381754] Fri, 21 March 2008 09:11 Go to previous message
Eclipse UserFriend
Originally posted by: mthusu.cs.helsinki.fi

Hello, yes this annotation thing sounds good. As KM3 is supposed to be
technologically independent and simple, PrimitiveTypes approach is not a
flaw at all.

-Regards Matti


Dimitrios Kolovos wrote:

> Hi Frédéric,

> Frédéric Jouault wrote:
>> Hello,
>>
>>> Have you tried Emfatic? It is quite similar to KM3 but EMF-specific.
>>
>> ... and therefore Java-specific.

> Indeed. However, I felt that practicality - as opposed to
> purity/technology independence - was the main concern in the original
> post, and Emfatic provides built-in support for the required features.
>>
>>>> I'm not sure what to say. This issue came up recently on the GMT
>>>> newgroup and it's something that KM3 folks need to consider.
>>>> Something like annotations to specify things not in KM3 but needed in
>>>> Ecore would seem to make sense.
>>
>> Indeed, this is how it works. Here is for instance the ATL annotation
>> model that complements the ATL KM3 metamodel:
>>
http://dev.eclipse.org/viewcvs/index.cgi/org.eclipse.m2m/org .eclipse.m2m.atl/dsls/ATL/Metamodel/ATL.ann?root=Modeling_Pr oject&view=markup
>>
>>
>>
>> Regards,
>>
>> Frédéric Jouault

> Greetings from York,
> Dimitrios
Re: km3 -> ecore - bad practice in metamodeling? [message #613646 is a reply to message #381751] Thu, 20 March 2008 19:06 Go to previous message
Dimitrios Kolovos is currently offline Dimitrios KolovosFriend
Messages: 1776
Registered: July 2009
Senior Member
Hi Matti,

Have you tried Emfatic? It is quite similar to KM3 but EMF-specific.

Cheers,
Dimitrios


Ed Merks wrote:
> Matti,
>
> I'm not sure what to say. This issue came up recently on the GMT
> newgroup and it's something that KM3 folks need to consider. Something
> like annotations to specify things not in KM3 but needed in Ecore would
> seem to make sense.
>
>
> Matti Husu wrote:
>> Hi, please tell me what do you think and is there a way to avoid the
>> problems below.
>>
>> First you specify metamodel with KM3 there has to be Primitive types
>> package if classes have attributes (String, Boolean...).
>>
>> Then you inject it to a .ecore file. Everything is cool. But if you
>> want to make an editor (EMF/GEF or GMF) you need .genmodel.
>> This can't be created at least on Eclipse 3.2.1/EMF 2.2.1
>>
>> In Eclipse 3.3.1.1 you can but you have to specify nsUri etc for the
>> PrimitiveTypes-package and corresponding java classes for Integer etc.
>>
>> I think GMF-ecore diagram editor is more handy.
>>
>> -Matti
>>
Re: km3 -> ecore - bad practice in metamodeling? [message #613649 is a reply to message #381752] Thu, 20 March 2008 21:45 Go to previous message
Frédéric Jouault is currently offline Frédéric JouaultFriend
Messages: 572
Registered: July 2009
Senior Member
Hello,

> Have you tried Emfatic? It is quite similar to KM3 but EMF-specific.

.... and therefore Java-specific.

>> I'm not sure what to say. This issue came up recently on the GMT
>> newgroup and it's something that KM3 folks need to consider.
>> Something like annotations to specify things not in KM3 but needed in
>> Ecore would seem to make sense.

Indeed, this is how it works. Here is for instance the ATL annotation
model that complements the ATL KM3 metamodel:
http://dev.eclipse.org/viewcvs/index.cgi/org.eclipse.m2m/org .eclipse.m2m.atl/dsls/ATL/Metamodel/ATL.ann?root=Modeling_Pr oject&view=markup


Regards,

Frédéric Jouault
Re: km3 -> ecore - bad practice in metamodeling? [message #614492 is a reply to message #381753] Thu, 20 March 2008 22:14 Go to previous message
Dimitrios Kolovos is currently offline Dimitrios KolovosFriend
Messages: 1776
Registered: July 2009
Senior Member
Hi Frédéric,

Frédéric Jouault wrote:
> Hello,
>
>> Have you tried Emfatic? It is quite similar to KM3 but EMF-specific.
>
> ... and therefore Java-specific.

Indeed. However, I felt that practicality - as opposed to
purity/technology independence - was the main concern in the original
post, and Emfatic provides built-in support for the required features.
>
>>> I'm not sure what to say. This issue came up recently on the GMT
>>> newgroup and it's something that KM3 folks need to consider.
>>> Something like annotations to specify things not in KM3 but needed in
>>> Ecore would seem to make sense.
>
> Indeed, this is how it works. Here is for instance the ATL annotation
> model that complements the ATL KM3 metamodel:
> http://dev.eclipse.org/viewcvs/index.cgi/org.eclipse.m2m/org .eclipse.m2m.atl/dsls/ATL/Metamodel/ATL.ann?root=Modeling_Pr oject&view=markup
>
>
>
> Regards,
>
> Frédéric Jouault

Greetings from York,
Dimitrios
Re: km3 -> ecore - bad practice in metamodeling? [message #614493 is a reply to message #381752] Fri, 21 March 2008 09:01 Go to previous message
Eclipse UserFriend
Originally posted by: mthusu.cs.helsinki.fi

Hi, not yet. I read about it and it's supposed to be good wih EMF. I will
try it. Right now I'm going to work with my thesis with the tools I have
chosen. I think I chose "KM3 over Emfatic" because KM3 is closely related
to ATL/AMW. For practical metamodeling Emfatic is maybe better. One way is
also to use GEMS. You get editor and all in just minutes. But... the
metamodel gets "polluted" with some mysterious stuff (Memento etc) and
diagram location stuff. I really don't understand why they haven't
separated this stuff from actual metamodel.

-Matti
Re: km3 -> ecore - bad practice in metamodeling? [message #614495 is a reply to message #381754] Fri, 21 March 2008 09:11 Go to previous message
Eclipse UserFriend
Originally posted by: mthusu.cs.helsinki.fi

Hello, yes this annotation thing sounds good. As KM3 is supposed to be
technologically independent and simple, PrimitiveTypes approach is not a
flaw at all.

-Regards Matti


Dimitrios Kolovos wrote:

> Hi Frédéric,

> Frédéric Jouault wrote:
>> Hello,
>>
>>> Have you tried Emfatic? It is quite similar to KM3 but EMF-specific.
>>
>> ... and therefore Java-specific.

> Indeed. However, I felt that practicality - as opposed to
> purity/technology independence - was the main concern in the original
> post, and Emfatic provides built-in support for the required features.
>>
>>>> I'm not sure what to say. This issue came up recently on the GMT
>>>> newgroup and it's something that KM3 folks need to consider.
>>>> Something like annotations to specify things not in KM3 but needed in
>>>> Ecore would seem to make sense.
>>
>> Indeed, this is how it works. Here is for instance the ATL annotation
>> model that complements the ATL KM3 metamodel:
>>
http://dev.eclipse.org/viewcvs/index.cgi/org.eclipse.m2m/org .eclipse.m2m.atl/dsls/ATL/Metamodel/ATL.ann?root=Modeling_Pr oject&view=markup
>>
>>
>>
>> Regards,
>>
>> Frédéric Jouault

> Greetings from York,
> Dimitrios
Previous Topic:Re: km3 -> ecore - bad practice in metamodeling?
Next Topic:Incubation, Parallel IP, & the GMT website
Goto Forum:
  


Current Time: Mon Dec 22 00:44:40 GMT 2014

Powered by FUDForum. Page generated in 0.02612 seconds
.:: Contact :: Home ::.

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