Hi Folks,
This article came up on the xml-dev list, so I read it and I am
passing it along.
http://www.xml.com/pub/a/2002/03/20/endpoints.html
I think that Rich's suggestions in the Embedded Use-Case for
wsia-specific terminology address some of the issues in the
typing/coupling choices we need to make, or for which we need to
provide. However, we have cases that need the entire range of strong
to weakly typed and tightly to loosely coupled AND static synchronous
(short, specific, immediate single-packet) to dynamic (long,
intermittent, streaming, changeable-stateful) messaging exchange
patterns. THAT's a lot to chew on. I suppose the question that is
nibbling at me is, do we need to chew on all of it at once?
We have broken our use-cases down (somewhat) into levels of
complexity. Should we start by working on the simpler LCD spec
requirements first and start writing spec on that while continuing to
explore the more complex issues? What concerns me is that we may get
rather further along in writing spec than the underlying structure
can support at present. I have a feeling that harvesting some
low-hanging fruit soon might influence the on-going debates about
HTTP, SOAP, etc. I have had the notion in my head for a while that we
might want to look at more than a single messaging transport
protocol, and see if such an arrangement would be amenable to
synchronization and integration with HTTP, SOAP, et al, and still be
deliverable in protocol/device independence.
Maybe I just like to worry, eh?
Ciao,
Rex
--