[
Date Prev][
Date Next][
Thread Prev][
Thread Next][
Date Index][
Thread Index]
[
List Home]
Re: [modeling-pmc] Re: [emft-dev] Proposal for Texo Component
|
Hi Sven,
I completely agree with your analysis about current generative
approaches in some common frameworks. I have considerable experience
with Seam. Seam contains a tool called seam-gen which generates a web
app on the basis of a set of annotated java classes or a database
schema. The issue I have with seam-gen is:
- the templates are difficult to customize
- no ide (==eclipse) integration (to accomplish for example code
formatting or project integration)
- no merging/re-generation support
- no (or very limited) way to influence the generation logic from the
outside (for example by extra annotations), or using a dsl
- the templates are just not advanced enough to cover many different
scenarios
- the generated application just does not look nice at all (this is a
personal opinion ofcourse, but imso generated apps should be demoable
right from the start)
When I then look at how EMF works and what EMF together with m2t tools
(workflow/xpand) can do then I am convinced (as you are) that there is a
large opportunity there for the Eclipse Modeling groups.
There is a strong tie between the proposed Texo project and EMF/EMFT,
Texo will be based on ecore-models and at runtime Texo will allow access
to EMF-models (from the generated code). I understand what you are
saying regarding domain-specific-applications. However because Texo has
a strong EMF relation, EMFT seems a good place to start. Depending on
the success and if there are other more logical main projects then it
can make sense to move there ofcourse.
gr. Martin
Sven Efftinge wrote:
Hi all,
this sounds very cool. We all know that there are many general
frameworks (web, RCP, SOA, etc,) out there which would profit from
generators.
Some of them (Ruby on Rails, Grails, SEAM just to name a few) already
use code generation a lot.
Unfortunately they use wizard-style passive code generation, where you
generate once and further change the generated code manually.
So most of the time generators are just used to get users up and
running as fast as possible.
Also they come up with concise languages to descibe domain models
(internal DSLs).
Those languages are not statically typed, have to fit into the host's
syntax and have no tool support at all (EMP can do better).
IMHO there should be some more projects like Texo, which show how EMP
can be used.
As EMFT is for EMF technology I find that those projects don't fit
well there. We should create a new Project for such domain-specific
applications of EMP.
What do you think?
atb,
Sven
btw: +1
On Apr 16, 2008, at 12:34 , Ed Merks wrote:
Extended EMF Community,
Martin Taal, the lead for the Teneo component, recently announced his
intent to work on the Texo component which will focus on exploiting
modeling technology for building web applications.
http://wiki.eclipse.org/Texo
I request the community's approval to create this new component as
part of
the EMFT subproject with Martin Taal as the component lead. Please
reply
with +1 or -1 on this thread. Obviously I recommend you give a +1,
given
Martin's outstanding track record with Teneo and his exemplary
support for
it. All EMF and EMFT committers are eligible to vote, so please take a
moment to do so.
Ed Merks/Toronto/IBM@IBMCA
mailto: merks@xxxxxxxxxx
905-413-3265 (t/l 313)
_______________________________________________
emft-dev mailing list
emft-dev@xxxxxxxxxxx
https://dev.eclipse.org/mailman/listinfo/emft-dev
--
With Regards, Martin Taal
Springsite/Elver.org
Office: Hardwareweg 4, 3821 BV Amersfoort
Postal: Nassaulaan 7, 3941 EC Doorn
The Netherlands
Tel: +31 (0)84 420 2397
Fax: +31 (0)84 225 9307
Mail: mtaal@xxxxxxxxxxxxxx - mtaal@xxxxxxxxx
Web: www.springsite.com - www.elver.org