RE: [i95, i22] - Proposal for clarifying use of SOAPAction

>>I would, however, like to discourage the use of SOAPAction for routing

>>of messages as this is not the place to put that information - the 
>>destination of an HTTP message is the request-URI.
>
>You say this a lot.  8-) And I don't quite understand why.

Thank you - I try to be consistent.

>It seems as though you're in favor of not restricting the 
>definition of SOAPAction or it's use and yet you say it 
>shouldn't be used for routing - it seems inconsistent.

What is inconsistent? HTTP request header fields are nothing new - they
compliment the request, as defined by HTTP [2]. Note that this has
nothing to do with SOAP or XMLP but rather a basic question of how HTTP
works.

Saying that it is ok to use SOAPAction for sending messages around
between HTTP applications makes no sense. What people might or might not
do is not under our control but we should at least make sure that we
don't encourage to do the wrong thing.

>Given the fact that some people are in fact using it for 
>routing, one of the problem I see with SOAP and the SOAPAction 
>field is that it introduces yet another place for specifying 
>what could be redundant information.  We have the request-URI, 
>the SOAPAction, and the TargetObjectURI on the SOAP Body 
>element - any of which could be used for 
>routing/dispatching/hinting.  No wonder there's so much 
>confusion and discussion.

Huh? We don't have a targetobjectURI anywhere? What's the confusion? 

Henrik

[2] http://www.w3.org/Protocols/rfc2616/rfc2616-sec5.html#sec5.3

Received on Tuesday, 8 May 2001 00:15:33 UTC