Binod,
You asked:
>>Even 236 doesn't necessarily need to define these interfaces
separately, right?
>>Given JSR 236 do have these interfaces defined separately, I thought
it should be
>>the decision of SIP Servlet EG to decide whether they should define
one
>>or more JNDI names. What I don't understand is how does it break, JSR
236.
No, JSR236 didn't have to separate those APIs but then again, neither
did JSE have to separate the ExecutorService and
ScheduledExecutorService. But it did and thus it was a natural
extension for us to do so also. Nathan's comment simply pointed out
that depending on the needs of SIP and the inheritance of those
interfaces, one "could" (vs "should") choose to register a single
interface in JNDI. In the end, it is the decision of the SIP EG to
make that decision.
>>I did understand the point you are making. Again, I thought, this
should be left
>>to the decision of SIP Servlet EG, unless it breaks 236 in some way.
One of the goals of JSR236 is to:
Allow application component providers to easily add concurrency to
existing Java EE applications.
Our point was that SIP could help meet that goal with the definition of
a default instance. It doesn't break JSR236 to not do so, and again it
is the decision of the SIP EG.
Regards,
Fred