application/rdf+xml type and assertions

At 15:50 18/03/04 -0500, Mark Baker wrote:
>Well, I am sort of, at least at this stage (i.e. prior to IETF last call);
>the application/rdf+xml draft says to send comments here.
>
>But certainly, if this goes last call without addressing this issue to my
>liking, then I'll mention that as a last call comment on ietf-types.

Mark,

With respect to your comments on the application/rdf+xml MIME type
registration, I note that we've exchanged many relevant ideas, but don't
seem to be moving toward any resolution.  Let me see I can summarize the
current position in terms that we can both agree:

- making assertions about the world is an important feature of the way we
expect RDF to be used, which sets it apart from many other data formats.

- this assertional _use_ of RDF is not fully covered by the current
published RDF specification, though the specifications do fully describe
the purely technical aspects of RDF assertions in terms of a formal model
theory [1].

- the RDFcore working group did try to address the matter of specifiying
that RDF documents are used to make assertions about the real world, but
were unable to achieve a widely accepted consensus, and did recommend that
a task force be set up to make further progress on it [2].

- a meeting of many interested parties [3] was held as part of the 2003
technical plenary, where it was agreed that the then-current text be removed
from the RDF technical specifications

- the Semantic Web coordination group did respond to RDFcore's request, and
have raised this as an issue with the TAG [4], and set up a task force and
mailing list to progress this matter [5][6].

I trust that there is nothing above with which you disagree, and that you
can see that RDFcore have devoted considerable attention to this issue.  By
way of a cross-check of some relevant facts, you may like to consult Dan
Connolly's summary to the TAG [8].


I would further make a couple of personal observations, but defer to the
Working Group chair or W3C staff contact for a definitive statement:

- the main purpose of a MIME type registration is to link the MIME type
name to its specification;  it is not the role of the registration to
override or modify the actual data format specification in any way.  As
such, registration is an administrative procedure, not a process of
technical specification.

- given the amount of attention that this matter has received through the
normal W3C process of creating the RDF technical specifications, I think
it's rather inappropriate to attempt to use the MIME type registration to
modify the working group's express decision.  I submit that the appropriate
way forward is to engage with the task force [5][6] that has been set up
for purpose of finding an acceptable resolution of this matter.


Accordingly, my recommendation to the RDFcore working group is to request
the IESG to approve the MIME type registration, along with informational
RFC publication of the registration template [7], for which I understand
that no IETF last call is required.

#g
--

[1] http://www.w3.org/TR/2004/REC-rdf-mt-20040210/
      http://www.w3.org/TR/2004/REC-rdf-mt-20040210/#intro

[2] http://www.w3.org/2000/03/rdf-tracking/#rdfms-assertion

[3] http://lists.w3.org/Archives/Public/www-rdf-interest/2003Feb/0073.html

[4] http://www.w3.org/2001/tag/issues.html?type=1#rdfURIMeaning-39

[5] http://lists.w3.org/Archives/Public/public-sw-meaning/2003Sep/0000.html

[6] http://lists.w3.org/Archives/Public/public-sw-meaning/2003Sep/0006.html

[7]
http://www.ietf.org/internet-drafts/draft-swartz-rdfcore-rdfxml-mediatype-04
.txt

[8] http://lists.w3.org/Archives/Public/www-tag/2003Jul/0127.html



------------
Graham Klyne
For email:
http://www.ninebynine.org/#Contact

Received on Monday, 22 March 2004 07:15:27 UTC