Re: Proposal: Uniqueness on the Wire Requirement for WSDL 2.0

FWIW, my position on this is that while there are advantages to putting
the operation in the message, I don't see that it has to be there.  So
long as the message contains sufficient information to identify the
requested operation, I think that should suffice.

An example would be a message sent to TCP port 999 meaning "store this
data" (there's no "STORE" operation in the message).  Instead of the
operation, port 999 would be associated - via public registration at
IANA - with a specification which describes the store semantic.

IMO, you have to take one of those two approaches.  The alternative, of
not having sufficient information in a message to understand it,
suffers from drastically reduced self-descriptiveness.

Mark Baker.   Ottawa, Ontario, CANADA.

Received on Monday, 27 October 2003 13:18:58 UTC