- From: Henrik Frystyk Nielsen <henrikn@microsoft.com>
- Date: Sat, 23 Feb 2002 11:00:49 -0800
- To: "Williams, Stuart" <skw@hplb.hpl.hp.com>, <noah_mendelsohn@us.ibm.com>
- Cc: <xml-dist-app@w3.org>
I agree with you analysis that a fault may not be sent anywhere, but it seems that there are many ways in which one can detect a failure in addition to receiving a concrete SOAP fault message. Regardless of how a fault is detected from the outside, I think it is important that we provide a deterministic state machine for a SOAP processor. Currently we provide two final states: "failure" and "non-failure" and as part of this I think we have to indicate how one gets to these states in all cases. Otherwise we end up with a funny third "don't know" state. Henrik >So, I don't think I regard a mechanism that merely generates faults, or >even is we go so far as to say (in accordance with the resolution of >Issue 102) that an MEP spec. MUST detail the disposition of faults >generated during the operation of the MEP, provides us with a reliable >mechanism to indicate failure. > >Even, if in all cases of failure we mandate the generation of a fault, >I think it would be a little unwise to *rely* on that as an "indication >of failure"... because there are no guarantees that it will be brought >to the attention of anyone/thing.
Received on Saturday, 23 February 2002 14:02:16 UTC