Re: [xml-dev] REST as RPC done right

[Paul Prescod <paul@prescod.net>]

Bill de hÓra wrote:
> 
> 
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> > -----Original Message-----
> > From: Simon St.Laurent [mailto:simonstl@simonstl.com]
> >
> > Intermediaries have a chance of understanding the wrappers.
> > We don't seem to write many intermediaries which have a
> > chance of understanding the contents, but that hardly means
> > it isn't possible, especially as the costs of processing
> > continue to drop.
> 
> Such intermediaries are fully possible and do exist. Cheaper
> processing is being amortized by XML data structures though.

An intermediary can deal with the data exactly to the extent that it
understands the message.

Barring AI, the intermediary understands the message to the extent that
the data's syntax and semantics are known in advance to the
intermediary, or at least the data can be mapped to something with
semantics the intermediary understands.

Teaching every intermediary about every bit of data in the world is not
cost effective. Therefore the obvious solution is standardization. We
need to make standards for the syntax and structure of the message.

Most messages consist of a request to send Data D to Process P or to get
Data named D from Process P. Therefore it makes sense to standardize the
method of naming processes (thus we have URIs) and the method of saying
"please get information" and "please send information". Thus we have
HTTP.

Of course you can build intermediaries with more detailed knowledge of
the structure of data D. But we can never completely standardize D
because some people need to send love letters and some people need to
send purchase orders. So we use XML for D because it is equally adept at
both.

Now given that we can't completely standardize D, IMHO we should at
least standardize what we can: the addressing model and the way of
saying whether you are trying to send information or get information.

 Paul Prescod