Re: [sca-j] NEW ISSUE: Redirecting callbacks

We will have fun debating the resolution to this.  I find the idea of a 
multi-hop forward to the real destination extremely unappealing. 
Presumably the reason for the current words was that the original caller 
may no longer be available, so this change would nullify the value of the 
capability.  Rather than reducing confusion, I think the change would 
increase confusion because the semantics in the case that the original 
caller has gone away would no longer be well-defined.  If we think that 
providing the currently specified facility is not desirable, I would 
prefer to simply remove the ability to specify a service reference as the 
callback object.

    Simon

Simon C. Nash, IBM Distinguished Engineer
Member of the IBM Academy of Technology
Tel. +44-1962-815156  Fax +44-1962-818999



"Michael Rowley" <> 
30/01/2008 20:13

To
"OASIS Java" <>
cc

Subject
[sca-j] NEW ISSUE: Redirecting callbacks






 
RAISER: Michael Rowley
 
TARGET: Java Common Annotations and APIs, section 6.7.5 Customizing the 
callback
 
DESCRIPTION:
 
The last paragraph of this section says:
 
A callback object may be a service reference to another service.  In that 
case, the callback messages go directly to the service that has been set 
as the callback.  If the callback object is not a service reference, then 
callback messages go to the client and are then routed to the specific 
instance that has been registered as the callback object.  However, if the 
callback interface has a stateless scope, then the callback object must be 
a service reference.
 
The second sentence of this (saying that ?... callback messages go 
directly...?) is overly constraining has complex and confusing 
implications for runtimes.
 
PROPOSAL:
 
Remove the second sentence from this paragraph.
 
 






Unless stated otherwise above:
IBM United Kingdom Limited - Registered in England and Wales with number 
741598. 
Registered office: PO Box 41, North Harbour, Portsmouth, Hampshire PO6 3AU