Re: Web-friendly SOAP

Jacek Kopecky wrote:
>...
>  If we agree on the high goal above, I believe we should not
> produce an HTTP binding of SOAP at all, or as a secondary binding
> for limited use in REST-based applications; the main binding
> being a transport protocol (TCP or other) binding.

We've agreed on this issue before Jacek. Running SOAP on TCP would
certainly be the more conventional and in many ways the more productive
way to go about this. But I'm not the person you must convince, that is
the others in xml-dist-app and the wider world. To be honest, I doubt
you will have much luck.

The point of us REST-ians is: "*If* you produce an HTTP binding it must
capture the semantics of HTTP as much as possible." That is an urgent
project because once the binding is deployed, it is deployed and if
people are abusing HTTP there will be damage to the Web infrastructure
as we saw with the Google example and its "missing URIs."

A less urgent project is to re-invent a REST-ful prototocol on top of
SOAP. In my opinion, this should have been a SOAP use case from the
start (as XHTML was an XML use-case). But it was not and now SOAP is not
particularly well suited as a protocol basis for the Web. In particular
it has poor handling for binary data. The web transports gigabytes of
non-Unicode data every day. Until SOAP has an efficient, standardized
solution for that issue, it is not an appropriate foundation for the Web
of the future. That is a pity.

 Paul Prescod

Received on Sunday, 16 June 2002 16:50:24 UTC