Re: [httpRange-14] What do HTTP URIs Identify?

On Thursday, August 8, 2002, at 08:36 AM, Bullard, Claude L (Len) wrote:

> Hmm.  It seems that all the identifier can do is
> identify the authority and the thing asserted by
> that authority.  This doesn't argue meaning other
> than THAT explicit source asserts THIS.
>
> That is ok.  It doesn't get around having a
> minimal two-level system with regards to
> interpretive meaning.  One cannot dispute that
> this authority says this, but one has to rely
> on
>
> 1.  Further assertions from that authority
> for interpretations
>
> 2.  Other assertions that point to that
> authority's assertions.
>
> Again, nothing wrong with that.
>


In fact, the web can produce both, so that a statement
can be completely "grounded in the web".

1.  If the assertions by the authority are made in RDF
using Properties which themselves have authoritative
information defining their meaning, then the first
authority does not have to and indeed may not
add extra information for the interpretation of his
statement.

2. (If I understand you right, you are saying
that we need some reason to respect that authority's
statements about the term.)
When an authority makes a statement at the URI
foo.rdf   about foo.rdf#bar, then the HTTP specs
are enough to say that that authority's word is
(bugs apart) definitive.  Any use of the #bar
identifier by anyone makes explicit reference to
the meaning as defined by that authority, and
therefore one needs no further pointers.

> I believe that we have come to some kind
> of consensus here that strongly suggests
> why the use of technologies such as RDDL
> and RDF have value:  to annotate and
> correlate the interpretations.  As has
> been stated, the use of the URI to point
> to a separate interpretive document is
> not mandated, but very useful.  The
> role of the URI in the XML namespace as
> a syntactical device to disambiguate names
> is fixed, not arguable.  The use of it
> to cite an interpretive document is not
> fixed, but it is wise.   To know what
> some authority asserts it means, we have
> to ask the cat.
>

Where does the cat come into it?
The authority can use RDF (OWL, etc)
to make assertions in terms of
other vocabularies. The authority
can use english, french etc.. to
assert those things which can't be done using
existing RDF vocabularies.

> The problem, of course, is to separate
> the ownership of the assertion from the
> ownership of the knowledge.  The Semantic
> Web will be a disaster if we can't do that.
> I don't think the solution is technical;
> it is legal.  It is also a different issue.
>
> len
>
> -----Original Message-----
> From: Bill de hOra [mailto:dehora@eircom.net]
>
>> From: Tim Berners-Lee [mailto:timbl@w3.org]
>>
>> However, on the web one *does* have a way to own and be the
>> authority on an identifier,  and there is no right of a third
>> party to argue that it means something else.
>
> ... Ok. What's good about that it squares with REST principles. While I
> don't see how we make people always do the right thing, I do see how we
> could build stuff on top of RDDL (or maybe DDDS) that could ask the
> authority directly for an answer to a questions about URI denotation in
> RDF.

Received on Monday, 12 August 2002 10:32:29 UTC