Dave Crocker wrote:
(appreciating the distraction from today's shock...)
At 08:45 AM 9/11/2001, Brian E Carpenter wrote:
There are other "SOAP over foo" proposals in the wings too. I do agree that
the general question needs looking at. Maybe it's OK to have lots of
SOAP mappings, or maybe it isn't. Although mtr may disagree, imho this does
need to be considered before deciding to publish a Proposed Standard.
Does it?
Your suggestion reflects a model that the IETF acts as a gatekeeper to
ensure only the right "number" of specifications.
I believe we have some responsibility to limit the number of
options. You may recall some roadkill called OSI, one of whose major
problems was too many options, notably at the transport level.
SOAP over HTTP
SOAP over HTTPR
SOAP over BEEP
SOAP over APEX
SOAP over TCP
SOAP over SCTP
...
ps. And of particular concern is that you are raising no specific
technical concerns, thereby leaving the matter as one that is purely abstract.
Putting the Web Services world in a situation where it needs a discovery
protocol
to find out how to layer SOAP is not at all an abstract issue. Or do you think
the WSDL for each Web Service is going to define the available layerings? What
is a poor vendor to do other than support all of them, and what is the technical
justification for that? imho this is an absolutely valid last call issue.
(If this had been a WG work item, it would have been discussed as a WG charter
issue.)
Brian