Jeff Bone jbone@jump.net
Thu, 17 Jan 2002 10:59:52 -0600

"S. Alexander Jacobson" wrote:

> I read some of the REST wiki and am seeing a lot
> of unactionable philosophy.   MIME-RPC seems to be
> consistent with the philosophy expressed there and
> to be actionable.
> If you could point out inconsistencies that might
> be helpful.

Hmm?  Couple of things:  first, I wasn't critiqueing MIME-RPC at all in this
thread, merely responding to your incorrect assertion that REST has no specs.  I'm
happy to take a look at MIME-RPC if you like --- haven't at all to this point, have
no opinion of it at all as it relates to REST.  I have no idea whether or not it's
inconsistent internally or with any other spec.

As for REST being "unactionable," it's not --- many existing Web systems constitute
existance proofs.  It really proposes nothing new that need be actionable.  REST is
an *architectural style* and the discussion of REST has been to date a discussion
of what characteristics of the Web give rise to what qualities.  I.e., it *is* an
engineering philosophy discussion.  The actionable underpinnings of REST are
largely (though perhaps not clearly enough) described in the existing
specifications for HTTP and URI.

That said, many of us recognize that it's a subtle discussion, and some clear
examples need to be offered.  To that end, the "DeRPC" / "RESTifying" bits on the
Wiki are a poor start;  Paul Prescod's article for XML.com (as yet unpublished
there, but draft is here [1]) is a great overview as well.



[1]  http://groups.yahoo.com/group/rest-discuss/message/18