Re: An analysis of mustUnderstand and related issues

Doug Davis writes: 

>> Just wondering, if we go with this 
>> approach, what does it mean if 
>> people don't use it? 

My suggestion would be, when you don't use the new feature, you get SOAP 
semantics.  Behavior would change only on headers where the new attributes 
were actually used:

* if a header says mustHappen, then when processed your header must be 
replaced by hasHappened

* if a header says dependsOn, then fault if precursors haven't happened

Other than that, not changes I think.  Keep in mind, this proposal is 
intended as a point of discussion.  I think it's worth considering, but we 
should compare it to whatever alternatives before becoming wedded to it. 

Even if we like the general idea, I think we should definitely explore the 
alternatives for exactly how we represent the dependency information in 
the message.  For example, I'm starting to change my mind:  maybe we can 
find a convention that's reasonably convenient and also models as a clean 
"module" (I.e. the existing mustUnderstand is used on some header to 
ensure that the new features are understood.)  I have some rough ideas on 
how to do this, which I will try to set down if I hit any breaks in the 
W3C schema meeting (where I am for the next two days.)  Thanks.

------------------------------------------------------------------------
Noah Mendelsohn                                    Voice: 1-617-693-4036
Lotus Development Corp.                            Fax: 1-617-693-8676
One Rogers Street
Cambridge, MA 02142
------------------------------------------------------------------------

Received on Tuesday, 15 May 2001 10:16:57 UTC