W3C home > Mailing lists > Public > www-rdf-logic@w3.org > January 2001

Re: Reification quoting in RDF/N3 was: A note comparing ConceptualGraphs and RDF/Semantic Web

From: Sean B. Palmer <sean@mysterylights.com>
Date: Sat, 20 Jan 2001 01:06:30 -0000
Message-ID: <002601c0827d$8995e980$8ca62cc3@z5n9x1>
To: "pat hayes" <phayes@ai.uwf.edu>
Cc: <www-rdf-logic@w3.org>
> >we can talk about properties of humans, and if we bring all of
> >these properties under one node, [...]
>
> That makes logical sense, but I don't honestly think it makes
> what might be called Web sense.

Only if you look upon nodes and ontologies as part of a hierarchy: but it
doesn't have to be that way. In other words, the data is just "there" as it
were; it's not necessarily definitive, it's just there to process if
required. Giving a set of data a URI that describes something that hasn't
got a URI is the only way I can see of grounding abstract facts in the Web.
It has to be decentralized: you can't have a registry of people, because it
has been proven time and time again that that wont work. I'm not saying
it's the perfect thing to do, I'm just saying it's the only way I can see.
Things on the SW that have URIs are easier to process...

> That is, I don't think that we can ever expect that all the
> information about anything is going to be gathered together
> in one place. (Who is "we" exactly, in your sentence
> above?)

But it is going to be gathered in one place: the Web. But it *is* a Web,
not a hierarchy, so it should remain decentralized. The "we" in "when we
want to re-use those properties" refers to anyone using the data that I
have put up. That would probably be something in my system, or based on my
system, or (hopefully) similar to my system. I can write a Schema showing
the equivalences and other ontologies of my properties, and as long as a
processor can find that (it will have a URL) and process it, then it can
work out what I'm saying about this particular person. This is all just
general SW stuff, really.

> Yes, but surely it doesnt have *all* the useful information about
> you; and, more to the point, what about all the people who don't have
> a home page? Surely we ought to be able to at least *refer* to them?

So then just write a load of information about that "person" and bung it in
the Web. Then you can refer to it...but you can't assert that it is
definitive. But then, that isn't essential in most cases, and even when it
is, we can rely on Digital Signatures. TmBL keeps saying that DS's will be
a backbone to the architecture of the SW, and it seems that every day I am
reminded of that fact by someone :-)

I'm sure another way to refer to "Pat Hayes" would be to drag out your
government records, SS number (or NI depending on where you come from), and
so on, but I think you'd prefer me to use your email address and a default
URL that you control? I'm pretty sure that I would, especially when the
data will be used on the Web itself, why not use Web axioms to tie that
data down? Something like:-

     "Sean B. Palmer" :hasHomepage <http://infomesh.net/sbp/sw/>;
     foaf:mbox <mailto:sean@mysterylights.com> .

is surely definitive enough, because it represents my interface to the Web
in the form of a unique mailbox (as defined by DanBri's foaf stuff), and my
SW homepage could contain my unique digital signature. Then, the data above
can be reused elsewhere, especially if I record in the Schema what the
various ontologies of my predicates are.

--
Kindest Regards,
Sean B. Palmer
@prefix : <http://infomesh.net/2001/01/n3terms/#> .
[ :name "Sean B. Palmer" ] has :homepage <http://infomesh.net/sbp/> .
Received on Friday, 19 January 2001 20:09:15 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:52:38 GMT