On 30 Nov 2005, at 20:04, Kenneth Tam wrote:
My take is that the existing codebases involved and the shorter/medium
term problems they're trying to solve are distinct enough that it
makes sense to let them evolve somewhat independently while
encouraging opportunistic integration..
Agreed.
having had some brief
exchanges w/ Dims & Sanjiva re: the relationship between this
proposal, Synapse, ServiceMix etc, we certainly plan to take advantage
of as many existing projects as possible. If given the opportunity
and some time, it turns out to make sense to most of the folks
involved that the Tuscany work ought to live under an existing PMC or
be rolled into an existing project, that's great.
Specifically, I think the goal of providing a language/platform
neutral service assembly layer that's backed by an extensible set of
app developer models (with a shared conceptual framework) for specific
languages/technologies, is something distinct here. Stuff like
figuring out how to work with e.g. ServiceMix being built around JBI,
a Java-centric standard, is definitely in the plan and seems like the
kind of work that would be appropriate to do while incubating.
Agreed. Though we can work on that after incubation too :)
FWIW we've been focussing quite hard lately in the ServiceMix project
on how to deploy different programming models into JBI; from SAAJ to
JCA to JAX-WS to vanilla JSR 181 POJOs; we're very keen to provide
all the JBI hooks so that SCA services (or a lightweight SCA
container) can drop right into ServiceMix/JBI so we can reuse the
various JBI services to work nicely with BPEL or smart routing
engines and so forth. Being able to reuse multiple transports &
routing & orchestration services can be quite handy as a deployment
option for SCA services.
I heartily support the Tuscany project ; its a good concise and
relatively small spec (I'm a big fan of small specs :). From a Java
perspective it certainly looks like Spring POJOs and SCA POJOs can
work nicely together with a very similar programming model; and JAX-
WS/JSR181 should be relatively easy to support too. SCA does seem to
be a competitor to EJB3 - I'm not the worlds biggest EJB3 fan so
competition in this area is a good thing IMHO - but I hope there
could be some close co-operation with the EJB folks in Geronimo as I
can see that despite its slightly different annotations & scope
models, the underlying implementation code for remoting, session
handling, load balancing and clustering will be very similar. e.g.
the Geronimo clustering work would be very useful to Tuscany.
In summary I'm very keen to see innovation in this area - its sorely
needed IMHO. I hope as Tuscany grows and matures it can present a
unified POJO programming model for Java developers that can work well
in other deployment scenarios (JAX-WS, pure Spring, maybe even EJB3
etc) to try simplify the application developers life somewhat.
Incidentally I've also wanted a nice Apache licensed SDO
implementation too for some time (thats not bound to EMF :)
James
-------
http://radio.weblogs.com/0112098/
___________________________________________________________
To help you stay safe and secure online, we've developed the all new Yahoo! Security Centre. http://uk.security.yahoo.com
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
