The What does SOAP really add? thread continues on xml-dev. Leigh Dodds has a good summary of the opening of the discussion, and it continues.

Particularly worth reading are posts from Don Box (1 2 ) and Adam Turoff, as well Julian Reschke's explanation of how SOAP is generally opaque to XSLT processing (1 2).

In the meantime, similar discussion is going on at the W3C's www-tag list. Anne Thomas Manes and Paul Prescod are going back and forth (1 2 3 4 5)

I'm wondering more and more if maybe it's time to sever Web Services from the Web. Practice, both technical and business, seems pretty different for the two sets of specifications. Don Box is already talking about moving away from HTTP, and the founding of the WSIO suggests that the corporate behemoths want their patents along with the Web Services.

It doesn't seem like it's worth risking the foundations of a successful communications medium just for the sake of envelopes embedded in documents and inflicting W3C XML Schema on the universe.

It's a risky proposition, but maybe the Web Services folks should take their toys and go play at the WSIO. There they can openly control the process, use RAND licensing, and migrate away from the Web as we know it. With that clear separation made, the Web can go on as a royalty-free mode of communications of a wide variety of sorts.