SOAPAction Proposal

All,
Right now the SOAPAction issue is a choice between:
  1 - Use of SOAPAction is discouraged. SOAPAction is an optional
      part of SOAP, supported but not required. Services MAY
      require SOAPAction and any software wishing to access those
      services MUST be able to send it.
  2 - Use of SOAPAction is deprecated. Senders SHOULD NOT send
      SOAPAction.  Receivers MUST NOT accept or reject messages
      on the basis of the presence, absence, or value of the
      SOAPAction header.

Looking at these two choices it seems like we really are actually
very close to an agreement.  IMO, they're actually (almost) the same.
Here's why: if we start with option #2 (aka kill SOAPAction) a Web
service can require any HTTP header be sent - and it is free to
reject the request if it is not sent.  So, given that, if the WG
decides that it doesn't like SOAPAction anymore and its use is
"deprecated" or "discouraged" then what we're really saying is
that we should say nothing about it - as if it never existed.  And,
if it never existed then people should be free to require/use any
application-defined HTTP they want - including one that just
happens to be named "SOAPAction".  This sounds a lot like option #1.
So I guess my formal proposal would be to have the spec say
something along the lines of:

  SOAP no longer requires the SOAPAction HTTP header.  While the
  definition of, or a Web services requirement of a SOAPAction header,
  as with any application-defined HTTP header, is left as an
  implementation choice and is outside the scope of this specification,
  a suggested usage of the SOAPAction header is that it can be used
  to indicate the "intent" of the XMLP/SOAP HTTP request.

(Editors will obviously need to fix it up, and we should probably
expand a little on what "intent" means but it's a start)
This seems like it should make the "kill SOAPAction" people happy
because it is no longer required, but should also make the people
who use SOAPAction (and see its value) happy because they can
still use it and provides for a "suggested" usage.
Thoughts?

-Dug

Received on Wednesday, 29 August 2001 11:12:29 UTC