Re: silly question about rdf:about

Hi Uche

> Hi, Nikita.  I must say that if anyone help make me understand this, you
> can.

Glad to hear that :-)

> > > And let me re-iterate that I disagree with this.  Eric Prud'hommeaux
said
> > the
> > > same thing, so I'm rather frightened a practice of this might be
forming.
> > >
> > > The straw man is that RDF uses http://uche.ogbuji.net to represent the
> > person
> > > "Uche Ogbuji".  All the discussion about "published subjects", and odd
> > tricks
> > > with unambiguousProperty seem to be solutions to this supposed
problem.
> > >
> > > But I don't see why anyone thinks that RDF says http://uche.ogbuji.net
> > *is*
> > > the person.
> >
> > If anybody thinks that way then it is not me.
> > What I am saying is that given a statement like this:
> >
> > http://uche.ogbuji.net :characteristics :very-interesting
> >
> > will leave many people puzzled whether I mean that you are a very
> > interesting person or
> > that your website is definitely worth looking at. :-)
>
> It shouldn't leave least bit puzzled.  It is obviously talking aboout the
Web
> site, not the person.
>
> > > If enough people agree that urn:folks:uche.ogbuji.net is an acceptable
> > > published subject identifier for "Uche Ogbuji", then they have already
> > done all
> > > the work RDF needs to take advantage of this in description of the
person.
> >
> > Well... what is the critical mass of "enough" ? When "enough" is enough?
>
> This depends on the use and the users.  In come cases, only one person
need be
> involved.  In others, all RDF users must agree (e.g. agreement on how to
treat
> rdfs:Resource).

Yes of course. This is the (TM) author's conscious decision. It is totally up to her/him.

> > > I don't see that Topic Maps gains anything with this built-in
indirection,
> > > except one of the most complex data models I have ever seen for a
> > description
> > > language (puts CIM to shame, I must say).
> >
> > The gain is in avoiding confusing situations like the one that I
mentioned
> > above.
>
> But I don't see the confusion.  RFC 1738, which governs the URI
> http://uche.ogbuji.net makes it clear that this URI locates/identifies the
> document that is retrieved using HTTP and that address.  Why would anyone
> thing it represents a person?

Actually, a correct statement would be something like:

One would think that I use http://uche.ogbuji.net to indicate a person 
if in my RDF Topic Map I have:

:Uche rtm:/indicatedBy http://uche.ogbuji.net; a :person.

> > I also do not see how this makes data model complex.
>
> Maybe I'm just thick, but I just do not come close to understanding Topic
> Maps.  There are just too many moving parts interacting in confusing ways.
I
> must say, though, from observice the discussiuons at KT, that I'm not sure
> anyone really does.
>
>
> > For one it allows equal use of more then one PSI for a topic so that
> > people speaking about a topic indicated by "urn:folks:uche.ogbuji.net"
> > and people speaking about a topic indicated by "http://uche.ogbuji.net"
will
> > understand each other.
> > And people speaking about 404 errors at "http://uche.ogbuji.net" will
not
> > interfere.
> > ( I ma getting a "Not Found" message)
> > :-))
>
> Really?  works from here, and http://validator.w3.org/ finds it OK as well
> (even though it finds it invalid: I'll look into this).

With IE6 I am getting redirected to 
http://uche.ogbuji.net:8080/uche.ogbuji.net/uche.ogbuji.net/main.doc?xslt=ma
in.xslt (see attachment).
Actually, with Netscape it works just fine (just tried it).

> > > This is *not* a flame on Topic Maps.  TM has things that RDF
desperately
> > > needs, such as scopes and merging, but I don't think that the
> > > subject/occurrence (or whatever) distinction is one of the things RDF
> > needs.
> >
> > Well there are other things too...
>
> Perhaps.  Anything in particular we should be discussing?
>

Well, one other thing that keep bothering me is that RDF uses XML IDs to 
create vocabulary entries. For this reason "Nikita Ogievetsky" is an invalid entry.
Topic Maps use baseName-s for that; and id attribute serves purely for addressing purposes.

(See http://www.cogx.com/xtm2rdf/extreme2001#slide54)


--Nikita

Received on Monday, 8 April 2002 01:06:37 UTC