Re: new issue: message passing a dual of shared state

Hi,

Some "conversational" systems, like those built with VoiceXML, reuse 
some of REST's constraints, but not all; they are explicitly stateful.
This isn't inherrently a bad thing, of course, because as has been
pointed out, voice apps *can* benefit from such an interaction model.
So you suck up statefulness, and take the bad (lowered reliability and
scalability) with the good (suitability).

If there's an issue buried here somewhere, it may be that it is unclear
to many to what extent the various constraints of REST are responsible
for the "Web-ishness" of some system (and inherit the extreme
interconnectivity of the Web, etc..), and at what point, while picking
and choosing between these constraints, would a system be considered not
part of the Web?  Or in other words, what is the minimum set of
architectural constraints that will make a system Web-like?

FWIW, my current view is that two constraints are critical; the uniform
interface, and identification of resources.  In this respect, I think
that VoiceXML systems are "on the Web" because they use both
constraints, whereas Web services are not on the Web, because they - at
least as currently implemented and specified - use neither.

MB
-- 
Mark Baker.   Ottawa, Ontario, CANADA.        http://www.markbaker.ca
Web architecture consulting, technical reports, evaluation & analysis

Received on Tuesday, 18 March 2003 07:21:01 UTC