Documents, Cars, Hills, and Valleys

[Diatribe on the range of HTTP and hash vs. slash issues. was: Re:
silly question about rdf:about]

> I agree with Uche on this one -- an http:// URI identifies
> a document, and not a car.  However, I can live with the
> other proposals that people have put forward (Sean's
> technique of adding an extra node works well).

Right. The problem with just declaring that the range of HTTP is a
document is that many eminent people will disagree; to name a few
names: Dan Connolly, Roy Fielding, Mark Baker, Dan Brickley, and Aaron
Swartz. Of course, there are many people on the document side too,
including Uche, TimBL, et al. (where "et al." probably includes most
users of the Web, to be quite honest). The point is that if the
argument were as clear cut as people on either side think it is, then
it would have been resolved by now.

From the RDF P.O.V. the argument is clearly important - it is one half
(or more) of the Hash vs. Slash issue, that I have been carefully
following for many months now. Once again, in this arena, some people
have been choosing hash (especially those who think that the range of
HTTP is a document), and others have been choosing slash (those who
think that HTTP URIs can identify anything). The hasists' views are
compatible with those of the slashists, except for those slashists
(e.g. Aaron Swartz) who would like to ban URIs+fragments from RDF
altogether.

Clearly, that's not going to happen in this version of RDF.
URIs+fragments have been deployed on a wide scale: the RDF "namespace"
itself contains a hash. RDF Schema and DAML contain hashes.

The main arguments for the range of HTTP being anything are:-

* It is minimally constraining, and that's good on a Web built to the
minimal constraint axiom (Dan Connolly used this tatic on www-tag).
cf. http://lists.w3.org/Archives/Public/www-tag/2002Mar/0188
* There is as yet no satisfactory definition of a "document" (Dan
Brickley raised this somewhere and when on www-rdf-interest). cf.
http://lists.w3.org/Archives/Public/www-rdf-interest/2001Nov/0174

I have tried to challenge the latter by coming up with a set of axioms
for what constitutes a document:-

http://lists.w3.org/Archives/Public/www-archive/2002Apr/0013

I urge anyone who believes in the range of HTTP being a document to
help formalize a definition of "document", otherwise you do not have a
strong case with which to back yourselves up. Documentists will also
have to argue with Aaron as to the utility of URIs+fragments in RDF.
In fact, I had a very interesting debate with Mark Baker on this
topic:-

http://ilrt.org/discovery/chatlogs/rdfig/2002-03-21.html#T03-50-04
http://ilrt.org/discovery/chatlogs/rdfig/2002-03-21.txt 03:50:04

To progress, we need to decide those two things: the range of HTTP,
and the nature of URIs+fragments. Note that if you (in general:
addressed to anyone who reads this) believe that the range of HTTP is
a document, and that hashes are unaccptable in RDF, then to you, RDF
is hopelessly broken and can never be fixed.

So that's a summary of where we're at right now. Something that is
frustrating for me is that people keep talking of the problem as if it
were just some philisophical excercise, without remembering that this
is a real-world problem that is making it difficult for people to
implement systems on top of RDF. It makes it difficult for me - since
it's difficult to know what any HTTP URI identifies, it's confounded
me in the EARL project, and it confounds me any time I have to select
whether a namespace ends in a hash or a slash.

The Semantic Web is turning out to be a looking-glass world, and
that's a shame on all those who have contributed to the architecture,
and a fear for all those who wish to contribute in future.

       "When you say 'hill,'" the Queen interrupted, "_I_ could show
     you hills, in comparison with which you'd call that a valley."

       "No, I shouldn't," said Alice, surprised into contradicting her
     at last:  "a hill *can't* be a valley, you know.  That would be
     nonsense--"

       The Red Queen shook her head, "You may call it 'nonsense' if
     you like," she said, "but *I've* heard nonsense, compared with
     which that would be as sensible as a dictionary!"

     - Through The Looking Glass, Lewis Carroll

--
Kindest Regards,
Sean B. Palmer
@prefix : <http://purl.org/net/swn#> .
:Sean :homepage <http://purl.org/net/sbp/> .

Received on Tuesday, 9 April 2002 07:42:30 UTC