[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]
Subject: RE: [wsrp] redirectURL allowed in handleEvents response?
I just want to observe that we can't really be tightly coupled as our explicitly distributed setting requires failures to be dealt with. A producer is not be able to tell the difference between a consumer that (silently) refuses to re-direct and a network error (loss of reply carrying redirect URL). Regards, Andre -----Original Message----- From: Michael Freedman [mailto:michael.freedman@oracle.com] Sent: 18 August 2005 18:55 To: sleland@notesdev.ibm.com Cc: wsrp@lists.oasis-open.org Subject: Re: [wsrp] redirectURL allowed in handleEvents response? PBI isn't optional from either the producer or the consumer perspective. Supporting events is optional from the consumer perspective. The producer must provide a signature for HandleEvents but can ignore any submissions. On your second point given technology we can't say its true that a HandleEvents would never need/want to do a redirect. My points however say that given the current limited semantics of redirect that in a loosely coupled coordination environment such a redirect has ambiguous meaning which if we leave open ended for the consumer to disambiguate leads to a situation where the consumer can operate the producer in an indeterminate state -- which I further go on to say is bad/wrong. In the tightly couple coordination environment where I believe its less ambiguous/more useful to redirect, I point out that the constituents of the coordination can rely on their tight coupling [i.e. they can use the event system itself] to send such redirects. -Mike-
[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]