W3C home > Mailing lists > Public > www-archive@w3.org > November 2009

FOAF and formalism

From: Dan Brickley <danbri@danbri.org>
Date: Thu, 19 Nov 2009 22:47:38 +0100
Message-ID: <eb19f3360911191347l72fa35e5x87aebf97201456c6@mail.gmail.com>
To: Pat Hayes <phayes@ihmc.us>
Cc: Thomas Baker <tbaker@tbaker.de>, www-archive@w3.org
Hi there Pat,

I was reponding in a Skype chat to your comments re FOAF and
formalism, thought I'd bounce it your way too...

The odd thing with thinking of  FOAF as an ontological 'specification
of a conceptualization', is that the underlying conceptualization of
people and their doings behind FOAF informs more the stuff I left out,
rather than the stuff I put in. And for example it's also visible in
the way we focussed on allowing identification-by-description (inverse
functional property reasoning) rather than forcing everyone to use
URIs "for people". There's certainly nothing much by way of classes
whose membership has nice tidy membership rules; except the odd
construct I put in for annotating a "Group" (or people/agents) with a
link to a class definition that can populate the group; so you might
have the group of W3C employees, defined as people whose
workplaceHomepage is http://www.w3.org/. That bit of the spec seems
almost entirely ignored so far, despite it being quite interesting!
Maybe not enough OWL tooling out there yet? (I was wondering about
putting in a mechanism for using SPARQL queries instead, ... but
forget that for now :)

Re social meaning, the best example of that in FOAF is
'schoolHomepage'. Where I come from, School is what you leave at 16 or
so. But it soon became clear that other English usage was dominating,
and 'school' was being used to cover universities etc too. It's
possible that having a richer ontology of educational institutions in
the spec might have stopped this happening, but I doubt it. Everything
points to a simple observation: people rarely read the spec (even
though it is fairly helpful and detailed prose, compared to many RDFS
namespaces which contain just RDFS/OWL).

See http://xmlns.com/foaf/spec/#term_schoolHomepage for documentation
of this shift in meaning, "FOAF does not (currently) define a class
for 'School' (if it did, it would probably be as a sub-class of
foaf:Organization). The original application area for
foaf:schoolHomepage was for 'schools' in the British-English sense;
however American-English usage has dominated, and it is now perfectly
reasonable to describe Universities, Colleges and post-graduate study
using foaf:schoolHomepage. "



[11/19/09 9:11:49 PM] Thomas Baker: Did you see Pat's comment at
FOAF is a good example of a
useful ontology which is almost nonexistent as a formal specification,
and gets all its meaning from the way its terms are actually used. It
is a socially defined ontology rather than a logically defined one.
Another example is dublin core. The primary use of the formal axioms
in cases like this is to be a guide for users in how to apply the
vocabulary to instance data, rather than to support complex inferences.
[11/19/09 10:33:01 PM] Dan Brickley: yeah, he's right and he's wrong
[11/19/09 10:33:05 PM] Dan Brickley: i was thinking about this recently
[11/19/09 10:33:20 PM] Dan Brickley: actually there's a reasonable
amount of ontology in foaf
[11/19/09 10:33:25 PM] Dan Brickley: esp compared to classic DC
[11/19/09 10:33:39 PM] Dan Brickley: eg. the identification of people
via their properties aspect
[11/19/09 10:35:02 PM] Dan Brickley: but per
http://www-ksl.stanford.edu/kst/what-is-an-ontology.html "    An
ontology is a specification of a conceptualization. " ...
[11/19/09 10:36:10 PM] Dan Brickley: ...there is a definitely a
worldview that foaf embodies, but the class/property structure doesn't
specify as such
[11/19/09 10:36:46 PM] Dan Brickley: for example, i believe it's
better to show friendship than to assert it. So in FOAF we do *not*
have a strict taxonomy of friendship types; instead we encourage
people to document the 'evidence friendship leaves in the world...'
[11/19/09 10:37:27 PM] Dan Brickley: we don't have much in foaf for
defining classes by their members
[11/19/09 10:37:47 PM] Dan Brickley: except for the OWL piece of the
group mechanism (similar to AgentClass), which is largely unused
[11/19/09 10:38:06 PM] Dan Brickley: i'll send this to Pat, see what he says!
Received on Thursday, 19 November 2009 21:48:12 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 14:43:36 UTC