[topicmapmail] Identities and names (WAS - A somewhat new topic maps format)
Peter P. Jones
ppj@concept67.fsnet.co.uk
Fri, 08 Aug 2003 19:07:31 +0100
Hi Daniel,
Please note that in what follows I'm not picking sides, just trying
to evaluate matters, possibly erring on the side of caution.
On 8 Aug 2003 at 17:01, Daniel Rivers-Moore wrote:
> Peter
>
> I think you are underestimating the ability of the RDF community to
> slide between seeing the referent of a URI to be a resource on a
> computer system to which the URI resolves, and seeing it as a
> "resource" in a much broader sense, namely any thing or concept.
I wasn't aware that I was, but perhaps I didn't come across well. I
am presently reading the RDF Semantics WD (yes, I know I'm a late
starter) and evaluating as I go. Imho, the distinction that draft
spec makes between a name, as mere label or string, and its
denotation, does not make the error that you ascribe above. It also
forms the basis of the W3C's formal mathematical description of the
computational properties of the RDF graph so it's power should not be
underestimated.
RDF
> does not make the distinction that XTM makes between what the XTM 1.0
> Conceptual Model calls a Resource (something addressable on a computer
> system) and what it calls a "Non-addressable Subject". RDF uses the
> word 'resource' to cover both, and uses URIs as names for both.
That's the point I know Steve Pepper is making. I'm aware of that
distinction, and from reading the aforementioned draft spec I think
TimBL is well aware of it too. But I think it is a distinct issue
from the point I describe above which lies at the heart of the RDF
Semantics WD.
The W3C is deliberately placing the emphasis of the model, or The
Model, within the machine. E.g. when I create an object model for a
piece of software I model objects in the world as say, Java objects,
and their tie to the real world is only via their appearance (serving
some useful semantic role) via the interface to the user. The W3C, it
seems to me, are taking precisely this line as regards the notion of
resources, classes, ontologising, etc.
In that infosphere there is only one notion, that of resource, and it
is conceivably all that is needed to drive effective SW interaction
if enough descriptive information can be loaded into that infosphere
too.
Imho, it would be a bad idea to underestimate those aspects of the
W3C vision.
>
> So, rereading the passage you quote below, which says that "any
> assertion of a graph implicitly asserts that all the names in the
> graph actually refer to something in the world", my conclusion would
> be:
>
> "Beware of just making up URIs that don't *refer* to anything"
>
> rather than
>
> "Beware of just making up URIs that don't *resolve* to anything"
I agree, except that I think the RDF Semantics WD implicitly carries
in its formal semantic notion of truth _only_ resolution, if I
understand it correctly, and I was speaking to closet RDFers. (Sorry
folks, wrong list; mea culpa.)
So I agree that you are right if it is understood that the notion of
'resource' in RDF only goes so far.
>
>
> One of the problems RDF has is that it does not provide any algorithm
> for determining what the URI refers to (there is no "protocol for
> dereferencing" the URIs in an RDF triple).
I'm guessing that's a deliberate delegation of responsibility by the
W3C.
They *may* refer to the
> resource to which they also resolve (if there is one), or they may
> refer to something that resource (if there is one) has a close
> relationship with (e.g. a person whose email address is used as the
> referencing URI). On the other hand, if there is no resource to which
> they resolve, they may refer to some resource to which they *did*
> resolve at some time in the past, or some resource the author believed
> would shortly come into existence, or
Not so. The formal semantics of RDF suggest that an Error 404 would
simply result in that clump of assertions using that 'dead' URI
evaluating to false (presumably discounting them from use in some
context).
..., or ..., or any other thing
> in the world (including non-addressable "resources/subjects" like
> people, places, mountains or ideas) that the author chose to use that
> URI as a name for.
Yep. And as I said above, imho that approach should not be
underestimated.
>
> In this regard, an interesting discussion went on throughout July on
> the W3C Technical Architecture Group discussion forum, with the
> subject "Meaning of URIs in RDF documents"
> (http://lists.w3.org/Archives/Public/www-tag/2003Jul/0159.html). I'll
> just pull our one tiny snippet from this thread which, I think,
> illustrates the problem (for RDF):
>
> Tim Berners-Lee says:
> If you take the case of an identifier for pat hayes <phayes@ihmc.us>,
> for example, the non-logician would consider that it identified one
> person and get on with their lives
>
> And Pat Hayes replies:
> The logician can say that also: it is the assertion that a single
> person exists who has that name. But (1) that is not the same as
> saying that the name - all by itself - "identifies" a single person
> (or, well, maybe it is: but if so, then other things said about URIs
> and resources are wrong) and (2) in fact, they don't assume that and
> get on with their lives. Sometimes they assume it indicates a person,
> sometimes a mailbox, sometimes a computer: it depends on the context.
Pat is correct, except in the RDFosphere the machine doesn't care.
For the machine the email address is just the unique resource to
which other properties can be attached. (TimBL is clearly having a
bad brain day in the above.)
Much of the required contexts for usage/resolution/whatever could
then be supplied in the application layer over the graph, again
delegating responsibility to the bits that are supposed to care as an
intrinsic aspect of their design.
> Topic Maps have, I believe, made considerable steps beyond the
> situation RDF finds itself in in this regard, and the work Bernard,
> Murray, and others on this list are urging us to take forward, will
> hopefully make still further progress
Yes, subjectIndicatorRef is a must have.
I also believe that RDF's reification mechanism is dreadful to the
point of utter uselessness as things stand.
And then there's blank nodes, whether RDFS is necessary, where does
metadata belong etc, etc...
Atb,
--
Peter
>
> Best regards
>
> Daniel
>
> -----Original Message-----
> From: Peter P. Jones [mailto:ppj@concept67.fsnet.co.uk]
> Sent: 08 August 2003 09:58
> To: topicmapmail
> Subject: Re: [topicmapmail] Identities and names (WAS - A somewhat new
> topic maps format)
>
> On 7 Aug 2003 at 19:26, Peter P. Jones wrote:
>
> > On 7 Aug 2003 at 15:08, Murray Altheim wrote:
> >
> > [...]
> > >
> [...]
> >
> > (Not including Steve Pepper's argument for distinguishing subject
> > location from mere resource location where unaddressable
> > containments are an issue...) The only weak argument I can raise
> > about the RDF approach as yet is the following: a) The set of names
> > is infinite. b) The set of URIs (as a subset of names) is smaller
> > but also infinite. c) The set of resolvable URIs is likely to remain
> > countably finite. d) Therefore, given finite time and processing
> > power isn't it better to have someway of distinguishing the set of
> > resolvable addresses in advance? Problem: How do you know that a URI
> > is resolvable in advance? Answer: Indicate that it should be. It
> > will at least provide some help.
> >
> > I've many pages of spec to go yet though, so maybe there's something
> > covering that later on(?).
> Thought so. There's an interesting paragraph in Section 1.4 of the RDF
> Semantics WD. To quote:
>
> "If the vocabulary of an RDF graph contains URI references that are
> not in the vocabulary [[PPJ ed. note: think that should be 'set of
> resources' or 'universe']] of an interpretation I - that is, if I
> simply does not give a semantic value to some name that is used in the
> graph - then these truth conditions will always yield the value false
> for some triple in the graph, and hence for the graph itself. Turned
> around, this means that any assertion of a graph implicitly asserts
> that all the names in the graph actually refer to something in the
> world. ..."
>
> So beware of just making up URIs that don't resolve to anything.
>
> >
> > --
> > Peter
> >
> >
> >
> > _______________________________________________
> > topicmapmail mailing list
> > topicmapmail@infoloom.com
> > http://www.infoloom.com/mailman/listinfo/topicmapmail
> >
>
>
> _______________________________________________
> topicmapmail mailing list
> topicmapmail@infoloom.com
> http://www.infoloom.com/mailman/listinfo/topicmapmail
> _______________________________________________
> topicmapmail mailing list
> topicmapmail@infoloom.com
> http://www.infoloom.com/mailman/listinfo/topicmapmail
>