RE: SWAP Example

> The method is specified as part of the HTTP protocol.  It is actually
> the first characters of the first line of the header of the request.
> Most requests to a web server use the GET method, so the browser will
> send as the first line to the server something like:
> 
> GET /pub/doc/index.html HTTP/1.1 
> 
> For this reason the method does not need to be part of the XML body of
> the message.  We were fairly well persuaded by ietf members who worked
> on WebDAV that this was the right way to extend the HTTP protocol for
> use in things like SWAP.  By the way, the WebDAV spec goes into detail
> about the PROPFIND method you used as an example.

I don't think it is appropriate to extend HTTP for SWAP
in the same way that HTTP was extended for WebDAV. WebDAV is
a set of extensions to HTTP for Web distributed authoring and versioning,
and the extensions being supported are appropriate for things
which might otherwise be accessed via web protocols.

I think SWAP's situation is much more like IPP's, where the
services are being accessed over the Internet are not otherwise
Web resources that one would access using a web browser, and
that using a single generic POST method (going POSTal), and putting
the actual operation in the body, makes more sense.

This design choice is somewhat of a religious debate, unfortunately:
it's clear that the first-order functionality of a protocol can be
expressed either way, and that the design considerations are second-order,
yet the design choice has a major effect in the description of (although
not the implementation of) the protocol.

Larry
-- 
http://www.parc.xerox.com/masinter

Received on Wednesday, 2 December 1998 17:07:53 UTC