MTOM Issues 431 and 432

I took an action item on the July 2nd concall to summarize the
discussion of issues 431[1] and 432[2]

Issue 431

What are the semantics of optimized elements? Specifically, are
intermediaries required to preserve optimization/non-optimization of
element content?

I think ( and serveral people agreed with me ) that intermediaries can
change the set of 'optimized elements'. In MTOM terms this means that
the set of values in the OptimizationCandidates property may change from
node to node. I think this is coherent, and potentially useful. Two
possible scenarios spring to mind:

A)	Passage through a non-MTOM aware intermediary, or transmission
to a non-MTOM aware end-point. 

	An initial sender optimizes a message resulting in one or more
attachments. An intermeidary detects that the next downstream node does
not support MTOM and transmits the message using standard SOAP 1.2/XML
1.0 serialization.

B)	Passage over a communication channel with different
characteristics.

	This one is a bit more tenuous right now, but the idea was that
a message might arrive at an intermediary with one set of optimized
elements and due to the characteristics of the communications channel to
the next SOAP node, it might make sense to optimize more ( or less )
elements.

Issue 432

How does a binding determine which nodes to optimize?

At the momemnt the spec is fairly clear that a binding can use any
number of ways of determining which elements to optimize including phase
of the moon, size of the data, QNames of the elements. 

Another question that came up is how does a receiver determine which
elements were optimized which led to discussion of a receiver side
property 'OptimizedElements'. This may be worth recording as a separate
issue.

I think this summarizes where we got to, if others have more info,
please wade in :-)

Gudge

[1] http://www.w3.org/2000/xp/Group/xmlp-issues.html#x431
[2] http://www.w3.org/2000/xp/Group/xmlp-issues.html#x432

Received on Friday, 4 July 2003 16:48:54 UTC