My top two issues

Hi all,

In response to Tim's request for a list of some issues, I offer my
thoughts.  This is by no means an exhaustive list.  I've just picked
the ones that, to me, seem to be most urgent to resolve as they're
either a) topical, b) potential sources of great pain in the future,
or c) both 8-).  Luckily, I've only got two.

Issue; URIs versus URI references

This issue flares up every so often.  It involves disagreement about
whether it's ok or not for RDF to make assertions with URI references.
Given that the R in RDF stands for "resource", some feel that it should
be limited to using URIs, not URI references, as URI references are not
URIs and therefore do not identify resources.  Others don't see the
problem.  I personally find it of great concern that so much of W3C work
is using URI references instead of URIs, without considering the
implications.  XML Schema datatypes does not include a URI type; anyURI
is a URI reference.

Aaron wrote up a good summary of this issue at [1].

Issue; HTTP is not well understood

I assert that the entire Web Services phenomena exists because of a
misunderstanding about what the HTTP protocol is about (and a whole
lot of marketing $$$ 8-).  IMO, it is critical that the TAG document the
generality of its application semantics.  While it is true that there
are Web-friendly uses of SOAP (which I remain the most outspoken
proponent of in the XML Protocol working group), I've yet to see an
independant developer using it in this way.  The TAG could save a lot of
people a lot of time and frustration by explaining the role and
relevance of HTTP to the Web.

In response to a lot of Web Services propaganda that's been floating
around the industry for the past while, I took at stab at doing this by
writing up an Internet Draft called "An Abstract Model for HTTP Resource
State"[2], which takes the view of HTTP as a state manipulation


Mark Baker, Chief Science Officer, Planetfred, Inc.
Ottawa, Ontario, CANADA.

Received on Monday, 17 December 2001 17:23:26 UTC