Re: names, URIs and ontologies

At 05:36 PM 10/27/00 -0500, pat hayes wrote:
[...]
>Allowing public names seems to me to be a no-brainer.  They will be 
>extremely useful and they will not disturb or interfere with anyone who 
>doesn't want to use them. They will provide a way for genuine social uses 
>of 'web logic' to evolve naturally. I can see no cost to allowing them, 
>other than that they violate some kind of doctrine. Arbitrary doctrines 
>which make my life more complicated and awkward than it needs to be are a 
>prime target for being questioned and, if necessary, disobeyed.

I found your thoughts to be very illuminating, touching on a few issues 
I've been wrestling with, though I'm not yet sure where they lead.


In a recent message (I forget the context), Dan C cited a Bertrand Russel 
coinage of 'Definite Description' in 
<http://www.ditext.com/russell/russell.html>.  I don't know if applying 
that can help to overcome the difficulties of non-uniqueness of proper 
(public) names that you cite.


In the RDF-IG, we have had some discussion of "anonymous" resources, which 
I think has some distinct parallels of your proper names.  The desire to 
use a label without necessarily knowing the globally unique URI(s) to which 
it may be bound (or with which it may be equivalent).  Your view of proper 
names, with the implied existential quantification, seems to me to capture 
many of the issues here.  I think we need something that performs in RDF 
the role that pronouns do for human languages.


Based on exchanges I have seen, I think the currently perceived way to 
achieve the effect of proper names in RDF is to do something like this:
    [<resource>] --properName--> "Boston"
or
    [<resource>] --cityName--> "Boston"
or
    [<resource>] --usCityName--> "Boston"
(where usCityName may be a sub-property of cityName, in turn a sub-property 
of properName.
As presented, the proper name itself is a literal, so we don't (currently) 
have formal ways to constrain its value or properties.  I suppose the 
literal might be replaced by an RDF resource with (say) '--value--> 
"Boston"' and various other properties.

But I think we still need a way to identify the resource so described, even 
though we don't have a global name for it, and be able to make other 
statements about it, not necessarily in the same document.  This is where 
the discussion of "anonymous" resources comes from (by which we mean not so 
much anonymity, but ignorance of a global name).


A topic that particularly interests me is using Contexts (based on ideas of 
McCarthy and Guha) to help model complex real-world relationships using 
RDF.  You discussed proper names with common usage shared across a number 
(a family?) of ontologies.  It seems to me that Contexts are well suited to 
capture this kind of relationship:  in effect, a Context can define (among 
other things) bindings between proper names and globally unique URIs.  Not 
every proper name used needs to be so bound (i.e. some may be left as just 
existentially quantified).  Different contexts may use different bindings.


Finally, an anecdotal observation.  The IETF (Internet Engineering Task 
Force) is currently considering a document that attempts to describe what 
is the IETF 
<http://www.ietf.org/internet-drafts/draft-hoffman-what-is-ietf-04.txt>. 
The IETF is not an incorporated body.  It has no formal membership.  It's 
not clear whether it is uniquely defined other than by purely social 
convention.  The proposed document describes the IETF in terms of the 
functions it performs and the procedures that it follows.  Fred Baker, 
chair of the IETF, has commented that as described, there might be more 
than one entity that conforms to the description given, and is this 
something that can be allowed?

I wonder if, in the final analysis, when using proper names we simply have 
to follow a line you suggest in another message (22-Sep, "Random thoughts 
on web logic", this list) and treat the assumption that (say) your "Boston" 
is the same as my "Boston" as an "inference fuse", accepted until 
specifically refuted.

#g

------------------------------------------------------------
Graham Klyne                       Content Technologies Ltd.
Strategic Research              <http://www.mimesweeper.com>
<Graham.Klyne@mimesweeper.com>
------------------------------------------------------------

Received on Wednesday, 1 November 2000 12:51:10 UTC