On Thu, 6 Dec 2007, Dave Courtois wrote:
in the last couple year i develop an application name Figura,
a distributed Cad system. I made use of Corba for this, and i find Corba
pretty well design and easy to use. Recently i have todo a web 2.0
application at work. So i chose JavaEE for that purpose because the
compagnie impose Java as langage. The use of this platform became a
nigthmare very very fast. Java EE it's long to understand, hard to
debug, and a simple application take alot of time to program.
I fully agree with you. Many people think that CORBA and C++ are complex
and difficult to learn. If they fail with the development of distributed
applications, then they blame CORBA. IMHO CORBA is a very good tool, if
you know what you are doing. If you don't know, well, then you are lost
anyway, with CORBA, Web Services, whatever.
So i decide to made my own C++ framework... So where to begin, and why
c++... After a long reflexion about the poor and const of C++ and why
object framework like corba and DCOM/COM+ was made i realise that, there
alot of benefit to start whit CORBA!!
For new applications, I'd not use plain CORBA anymore. Here at
ObjectSecurity, we switched to the CORBA Components Model (CCM). At least
for implementing the application's core functionality. Plain CORBA is
mainly used for integration of legacy or 3rd party software which cannot
be turned into a component. CCM provides many very useful features for
complex distributed applications, e.g. setting up connections or
deployment.
The real fun starts when you combine CCM with a MDA tool chain. You just
model your system and press a button to generate IDL, CIDL, skeletons and
descriptors. Then you add your business logic. And that's it!
We are using Qedo (www.qedo.org), the CCM implementation on top of MICO,
for some very complex systems. Together with the ObjectWall IIOP proxy and
the OpenPMF Policy Management Framework we are also able to run
distributed systems securely over domain boundaries.
We used this platform e.g. for a C2 system and distributed airport/air
traffic simulations. It saved us a lot of development effort and
produced better (more flexible and more maintainable) applications.
The next project in the queue is a very complex and large transportation
system.
So before you think about implementing a new framework, have a look at
Qedo. It might save you a lot of work.
So i plan in the first part of that project to made nice implementation of
the pesitence state service, that can be use by MICO.
This would be most welcome!
Cheers,
Rudi
_______________________________________________
Mico-devel mailing list
Mico-devel@mico.org
http://www.mico.org/mailman/listinfo/mico-devel
