- From: Jacek Kopecky <jacek@systinet.com>
- Date: Wed, 30 Jan 2002 20:51:44 +0100 (CET)
- To: Mark Baker <distobj@acm.org>
- cc: <xml-dist-app@w3.org>
Mark, I'm not sure what you mean by "web architecture", because I thought we have yet to see that. 8-) Anyway, from SOAP-centric point of view, a transport binding specifies means of getting SOAP messages from one node to another. I think that your proposal (of making the Body optional) should come from a use case independent of any particular protocol binding, not from some unusual but possible use of GET method in HTTP. Do you have any use case for Body-less messages that is independent of HTTP? Best regards, Jacek Kopecky Senior Architect, Systinet (formerly Idoox) http://www.systinet.com/ On Wed, 30 Jan 2002, Mark Baker wrote: > Hi all, > > I had a thought recently, about an extension to our SOAP/HTTP binding > to support the GET method. I'm *not* suggesting we incorporate this > into the spec, only that we consider its needs going forward, as it > requires a change to the current envelope schema. > > AFAIK, no use of HTTP GET today includes a body, though RFC 2616 doesn't > preclude one from being used. But, it defines no behaviour for what the > body should mean, except implicitly in the definition of GET that it > can't change the meaning of the GET (i.e. can't use the body to change > what the Request-URI identifies). > > So, if we used the GET body to carry header information, specifically > SOAP headers, we could use GET in a web architecture friendly manner. > What that means is that a GET binding could permit an envelope in a GET > body, but that envelope would have to be all headers, no body. However, > currently, the envelope schema requires that there be at least one body. > > Would we consider making this change to the envelope to support GET in > the future? > > Note that this may help address issue 133. > > MB >
Received on Wednesday, 30 January 2002 14:51:47 UTC