|
What format should the source and target models be?
|
|
Replies:
1
-
Pages:
1
-
Last Post:
Jul 9, 2008 10:42 AM
Last Post By: KongFu-Panda
|
Threads:
[
Previous
|
Next
]
|
|
Posts:
1
Registered:
Jun 05, 2008 09:24:53 AM
|
|
|
|
What format should the source and target models be?
Posted:
Jun 05, 2008 09:51:01 AM
|
|
|
|
Can someone clarify the following for me.
Ultimately what I want to do is write a transformation between a source metamodel and a target metamodel (both expressed in ecore), specify a source model as an instance of the ecore source metamodel (using xmi) and run the transformation to give a target model expressed as an instance of the ecore target metamodel in xmi format.
I'm looking at the Java2UML sample that is bundled with the MTF eclipse plugin. The transformation refers to elements in the Java and UML metamodels and to run the transformation you give it a Java program and it returns a .uml file.
My question is how is the Java program converted to an instance of the Java metamodel and similarly how is the generated instance of the UML metamodel converted to the .uml file? How do you specify the format in which the source and target model are to be supplied to the transformation i.e in this case a Java program and .uml file rather than xmi instances of the ecore metamodels?
Is this something built in to MTF? What is going on behind the scenes?
Thanks.
|
|
Posts:
1
Registered:
Jul 09, 2008 10:25:09 AM
|
|
|
|
Re: What format should the source and target models be?
Posted:
Jul 09, 2008 10:42:45 AM
in response to: jmcq's post
|
|
|
|
hi,
without knowing the meta models of source and target models, the transformation between which would be so hard. For your case Jave2UML, MTF knows how Java Programme should be written(meta model of java language)and the metamodel of EMF models (EMOF?ecore metamodel?i am not sure which is subset of MOF).
By the way MTF use mapping its own language, not QVT, which i've just knowen.
best wish
|
|
|
|
|
