Ok it strikes me that we have something quite powerful on our hands and we
should really document it well for the FINAL release (soon now).  BTW, the
current authors of JBoss Group are coming out with the free version of the
doco so that the guys on the dev list can see what has been done and we will
put a todo out so you can start helping as you want.  But the point is that
specifying a client and server stack is what we do today and that specifies
the EJB behavior.  I find that kind of interesting it means we are already
beyond the EJB definition by making it all standalone and assembling it,
within the MBean framework, to give the EJB behavior.

I want to specify clearly the JMX behavior because we will end up with an
abstracted superset to the EJB specification.

I believe it is going to be also a more complete implementation than JSR160
(on which 2 of JBoss Group people are, juha and andreas) which deals with
next generation JMX.

The only things that are missing are
1- a persistence definition.  Frankly EJB 2.0 IS THE SILVER BULLET, they got
it right
2- a comprehensive discovery mechanism, EJB's is powerful with a home and
therefore access to a cache (findByPrimaryKey)
3- A CACHE STRUCTURE.

I believe we could open our MBean lookup to more advanced lookups than JNDI
and specify a generic persistence mech. Model Mbeans are interesting in the
persistence I believe, could be wrong

Just random R&D thoughts on a Sunday morning

</sunday-mass>

marcf


xxxxxxxxxxxxxxxxx
Marc Fleury, Ph.D
President
JBoss Group, LLC
xxxxxxxxxxxxxxxxx



_______________________________________________
Jboss-development mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/jboss-development

Reply via email to