RE: [KEYPROV] Re: [APPS-REVIEW] Review of HTTP Binding for DSKPP

["Hallam-Baker, Phillip" <pbaker at verisign.com>]

> From: Mark Nottingham [mailto:mnot at yahoo-inc.com] 
> Without speaking to BCP56, a few things here caught my eye...
> 
> On 2007/10/05, at 5:11 AM, Hallam-Baker, Phillip wrote:

> > It is not helpful here if the IETF is going to insist on a separate 
> > definition of Web Services specifications that is not in sync with 
> > where the Web Services world is.
> 
> I'm sure both the W3C and OASIS would love it if the locus of 
> the "Web Services world" could be nailed down, never mind the IETF.

Yes, there is certainly an element of the old VMS manual problem. If you remember the days of the grey wall, there were 50 or so binders which each described one aspect of the VMS architecture in mind numbing detail. Most people simply gave up.

If you looked long enough you might have come across a volume called the VMS software architecture which was the key to the whole manual set. The architecture manual explained all the things you needed to know like how DCL worked, how AST completions worked and the like. Unless you happened to strike lucky and find that manual there was simply too much to absorb.

Web Services is worse because there isn't a manual.


> > Either the BCP56 view is right in which case we need the 
> proponents of 
> > this view to be talking to the wider Web Services world (OASIS,
> > W3C) and arriving at a consensus position or the BCP56 view is 
> > obsolete and needs to be updated.
> 
> The W3C, for one, has a very clear position about the 
> relationship between the Web and Web Services, which includes 
> use of HTTP; see documents by the TAG, for example.

The IETF needs to be saying the same thing, not something different.


> > The port number issue is somewhat more complex. The number of Web 
> > Services is rapidly expanding and the idea of one port per 
> Web Service 
> > is simply not sustainable. We only have 65536 ports and we 
> are going 
> > to have far more Web Services in use.
> 
> If you expand "Web Services" to include what some people are 
> now calling "HTTP Web Services" or "RESTful Web Services," I 
> agree that
> BCP56 needs some clarification.

Well I don't use the term RESTfull because I don't think those following it are really follwing Fielding's architecture so much as they are doing Web Services over raw HTTP rather than with SOAP.

That is also fine and in fact that is why I think discussion of Web Services should begin with WSDL rather than with SOAP. You can do a Web Service with or without SOAP and you can do a Web Service with or without XML schema. But if you end up with something that you can't describe in WSDL it probably isn't a Web Service.


_______________________________________________
APPS-REVIEW mailing list
APPS-REVIEW at ietf.org
https://www1.ietf.org/mailman/listinfo/apps-review