- From: Greg Stein <gstein@lyra.org>
- Date: Thu, 04 Mar 1999 02:15:10 -0800
- To: "Roy T. Fielding" <fielding@kiwi.ics.uci.edu>
- CC: jamsden@us.ibm.com, w3c-dist-auth@w3.org, ietf-dav-versioning@w3.org
Umm... Roy, I think his note was calling for a way to say "I defined a book author property using <this> DTD and semantic" so that others could use the same specification with the intent of some level of interoperability. I didn't read anything in there about HTTP methods or other protocol-level changes. As an empirical example, Lisa Lippert's draft spec for additional collection properties is a great case-study for publishing some "standardized" properties [that occur outside of the needs of the base protocol(s)]. Cheers, -g Roy T. Fielding wrote: > > >One approach is to take a lesson from XML. XML is a standard language for > >tagged documents. ... > > I don't want to burst any bubble here, but XML has yet to become a lesson > for anything but pre-release marketing hype. What you say is true of any > codification of application domains, and metadata efforts in general. > If we understand and agree upon the key elements of an application > architecture, then standard identifiers of those elements can be used > to construct self-descriptive actions. That is a good thing in any syntax, > but something that the IETF usually avoids doing within a WG. > > The problem with using that style to define a protocol, however, is that > all pieces of the communication apparatus cannot be expected to understand > all application domains, and certainly cannot be expected to download > application domain descriptions. There must be a standard set of general > semantics that is sufficient to identify what type of action is taking place, > not the details of how it will be implemented. That is what HTTP provides > with the abstract resource interface, of which WebDAV is now a part. > > As I said a while back, if a specific application is calling for a > large number of HTTP methods to be defined, then that application is > misusing HTTP. Almost always, the solution is to rethink the role of > resources in the application. > > ....Roy -- Greg Stein, http://www.lyra.org/
Received on Thursday, 4 March 1999 05:18:04 UTC