[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[news.eclipse.technology.omelet] Re: Hi from OMELET

Hi Ed:

> No. OMELET has no internal transformation technology - the only
> one it might have is that of composing sub-transformations all of which
must
> be in an added-on technology. So OMELET has add-on support for
> Java/XSLT/ANT transformations with Tefkat and ATL in progress.

This is the normal way of cross-using contributions from one Eclipse project
to another one. GMT is essentially a repository for MDD research
prototypes and several new tools/resources are going to be commited
in the next weeks/months. In GMT there may be several concurrent
transformation technologies for example. The integration facilities
defined in MDDi will allow these tools to interoperate in stream mode (XMI?)
or in event mode for example.

> There are many interesting transformation related tools that would benefit
> from integration with a/the ModelBus. However it is not possible for the
> core team to understand all of them - I found it very distracting spending
> a little time looking at X then Y then Z and wanting to integrate all of
them.
> You need a small (perhaps 5) very committed (almost full-time) team to
> establish the ModelBus APIs/protocols. Then the many other enthusiastic
> contributors can provide the appropriate integration of the tools and
> technologies with which they are familiar.

I agree with your vision on this. Your idea of types (metamodels)
is a good basis to start with. Adding event-based and service-based
capabilities is also important if we are able to cope with all that
in a regular integration framework that will stay conceptually simple.
(We do not need to reinvent PCTE at this point). IMHO this is
the real challenge of the MDDi proposal.

> GMT and OMELET separated into two different projects because there
> were two different approaches. GMT seeks to provide a good user-oriented
> MDA capability. OMELET (like MDDi) sought to provide the low-level
> integration of whatever tools are available.
> ATL is a good QVT prototype that should plug-in to a/the ModelBus.

The plans are to make adaptors for ATL available in the MDDi project,
ATL itself and other prototypes being contributed and developing
in the GMT project. As soon as the MDDi integration specifications
are ready, complete and stable, this work is going to be undertaken
by one of the ModelWare partners.

> I am not sure that GMT has any other existence than a facade for
> these two modelling sub-projects.

>From the external state of the project you may be right.
However again there are a number of new MDD prototypes that
have been recently proposed and that are going to broaden the
spectrum of GMT facilities.

> I think that GMF and MDDi should provide the basis for the high level
> integration that GMT might be after.

I completely share this vision.
MDDi is going to provide the integration facilities
to many various functional blocks that are going to be contributed in
other projects.

Jean

________________________________________

Jean Bézivin - ATLAS Group (INRIA & LINA) - University of Nantes - 2, rue de
la Houssinière - 44322 Nantes cedex 3 - France

tel. +33 2 51 12 58 13 - fax. +33 2 51 12 58 12 - cell.+33 6 14 32 22 36 -
new e.mail: bezivin@xxxxxxxxxxxxxx

http://www.sciences.univ-nantes.fr/lina/atl/