RE: on documents and terms [was: RE: [WNET] new proposal WNURIsand related issues]

Dan,

> From: Dan Connolly [mailto:connolly@w3.org] 
> 
> On Fri, 2006-05-12 at 01:54 -0400, Booth, David (HP Software - Boston)
> wrote:
> > Dan,
> > 
> > Thanks for the very helpful explanations, and sorry it's 
> > taken a while to respond.
> > 
> > When the WebArch is unclear (or even conflicting), readers have no 
> > choice but to guess the TAG's intent.
> 
> What I'm trying to explain is that the TAG's position isn't 
> conflicting nor unclear; 

It's no use claiming that the TAG's position is clear if other people
find it unclear.  In spite of the WebArch being a superb work and I
think mostly clear, I, for one, find it unclear on this particular
question.  I have been unable to figure out the exact distinction
between an "information resource" and a non-"information resource".  I
*thought* I understood, but I was admittedly guessing, and it sounds
like you think I guessed wrong.

> it just says less than people in 
> this group have claimed.

I think saying too little makes it impossible for the TAG's guidance on
this point to be clear.  The httpRange-14 decision defines a rule for
people to follow.  The antecedent of that rule is whether a resource is
an "information resource".  If TAG says too little about what is the
distinction between "information resource" and non-"information
resoruce", then it is impossible for the rule to be clear.  This has a
very practical impact: people are earnestly trying to follow the TAG's
guidance in properly minting their URIs and serving associated metadata,
and it isn't clear whether the rule applies.  

The interpretation I proposed is conservative in that, unless the TAG
outright reverses itself, users are sure to conform to the TAG's rule,
even if the TAG later indicates that some previously uncategorized
resources are really non-"information resources".

> 
> >  You've given one interpretation,
> > which sounds like it boils down to:
> > 
> > 	- Any resource r is an "information resource" if:
> > 
> > 		a. there exists a URI that returns a 
> > 		2xx status when dereferenced; and
> >
> > 		b. r is not a dog, person or physical book.
> > 
> > 		c. the owner of that URI claims that the
> > 		the URI identifies r; and
> 
> I don't know how you got that from what I wrote. 

Here's how.  In [20] you say:

>  _:reply1 a http:OKReponse;
>   http:about <http://dm93.org/2006/05/05/abc> mime:body "abc";
>   mime:content-type "text/plain".

I presume that an http:OKResponse comes from dereferencing a URI that
exists.  Therefore I got the following condition:

> > 		a. there exists a URI that returns a 
> > 		2xx status when dereferenced; and

In that same message you also said:

> To re-iterate, I'm attributing 2 RDF/turtle formulas to the TAG:
> 
> 1.  w:representation rdfs:domain w:InformationResource.
. . .
> 
> 2.  w:InformationResource owl:disjointFrom foaf:Person.

from which I got:

> > 		b. r is not a dog, person or physical book.

(Well, I also added dog and book that you mentioned elsewhere.)
And in the same message you said:

> Further, I claim that
>   <http://dm93.org/2006geo/abc> owl:sameAs _:repr1.
> 
> i.e. not only does it have a representation that is "abc",
> it is _identical_ to "abc".
> 

and since you seem to imply that there is significance to the fact that
you are the owner of the URI when you make this claim:

> This claim is (a) mine to make, as owner of dm93.org,
> and (b) logically consistent with the position of the TAG.

(i.e., this claim has greater significance than any other random claim
than anyone might make) it sounded to me as though you were suggesting
that the URI owner's *claim* about what a particular URI names, is
significant in whether the WebArch considers that resource to actually
*be* what the owner claimed.

Thus I got:

> > 		c. the owner of that URI claims that the
> > 		the URI identifies r;

However, I'm now guessing that what you meant to illustrate was merely
that the interpretation of the TAG's position that I offered does not
*necessarily* follow from the TAG's writings, because another less
constraining interpretation, which conflicts with the interpretation
that I offered, is also logically consistent with the TAG's writings.
And I agree with this.  As I said, the interpretation that I offered was
an attempt to guess the TAG's intent.

> Perhaps
> I should just stop. The first condition is sufficient.
> 
> If it looks like I'm trying to say anything more than
> what the TAG has published, I really should just stop.
> I'm really, really trying not to.

My problem is that I don't know how to interpret some details of what
the TAG has written.  Example:

> "URI ownership gives the relevant social entity certain rights,
> including:
>      2. to associate a resource with an owned URI"
>   -- 2.2.2. URI allocation
>   http://www.w3.org/TR/2004/REC-webarch-20041215/#uri-assignment

I had previously taken this to mean that the URI owner has the right to
be distinguished among all others who may make claims about what that
URI names, but that this right does *not* imply that the URI owner's
claim is necessarily true according to the WebArch.  However, in the
absence of better factual evidence, if a URI owner claims that the URI
identifies a particular resource, then we might as well believe it
unless and until we find actual evidence to the contrary.  Example: if
Mark Baker asserts that http://markbaker.ca/ identifies Mark Baker the
person, then we might as well believe in until we notice that
dereferencing the URI yields a 2xx response, in which case we then
*know* that, according to the WebArch, Mark Baker's assertion is false.

However, your http://dm93.org/2006geo/abc example [20] has made me
realize that there is a different potential interpretation of the above
WebArch statement.  In particular, it can be interpreted as saying that
if a URI owner claims that a URI names a particular resource, then that
URI *does* name that resource (according to the WebArch).  This
interpretation would imply that the TAG has given two different rules
for determining what a resource names, and I have not seen any guidance
about which one should take precedence if they are in conflict, such as
in the http://markbaker.ca/ example.

I'm sorry if this is exasperating.  I'm really trying to figure this
out.

[20]
http://lists.w3.org/Archives/Public/public-swbp-wg/2006May/0035.html

David Booth

Received on Wednesday, 17 May 2006 09:38:01 UTC