[graap-wg] minutes from 4/24 telecon

[Karl Czajkowski]

On Apr 24, Heiko Ludwig modulated:
> 
> The ContinuingFaultType is the specific WS-Agreement fault that we
> defined. I don't know anymore the rationale why it was called that
> name, but we might want to have a sepate fault type.
> 
> Karl, do you rember the semantics of this name?
> 
> Heiko
> 

Yes, I think it is obsolete.  It had to do with distinguishing
continuing or terminal faults in negotiations, i.e. continuing fault
is like an E_BUSY etc, temorary failure which cancels an operation
while terminal fault would be like a permanent fault on the
negotiation (web resource) meaning the resource is no longer useful.
I think this is much less important with the current offer/accept
handshake, i.e. there is no ongoing conversation of counter-offers
that you might or might not want to terminate.

For Agreement resources, I think we can leave it up to the individual
fault types to define their semantics, and not bother with this.  For
example, it is going to be underlying WS-Addressing faults that
indicate your agreement resource EPR is invalid, right?  We don't even
specify the important faults.

karl

-- 
Karl Czajkowski
karlcz at univa.com