W3C home > Mailing lists > Public > public-esw-thes@w3.org > September 2004

Re: [Proposal][SKOS-Core] skos:denotes

From: Dan Brickley <danbri@w3.org>
Date: Wed, 29 Sep 2004 14:41:24 -0400
To: Stella Dextre Clarke <sdclarke@lukehouse.demon.co.uk>
Cc: public-esw-thes@w3.org, "'Miles, AJ (Alistair) '" <A.J.Miles@rl.ac.uk>
Message-ID: <20040929184124.GW2012@homer.w3.org>

Hi Stella,

* Stella Dextre Clarke <sdclarke@lukehouse.demon.co.uk> [2004-09-29 18:20+0100]
> Sorry, but I am completely lost in this conversation. Seems to me what
> this bloke Alistair J Miles is really looking for is a sort of "Beam me
> up Scotty" over the Internet. (Any minute now he'll pop right out of my
> monitor.) I'm not just making a joke of it. I think I am saying you are
> asking for the moon, and any approximation to the moon could prove
> unsatisfactory.

It's a tricky idea to explain, but I've seen enough RDF applications
that describe people and other entities directly that I know it's not a
moon-on-stick. So I think the problem is with communicating the basic idea 

Let me try to redescribe the issue...

SKOS uses W3C's Resource Description Framework (RDF) to describe
thesauri (and similar things).

Other projects (eg. FOAF, RSS, MusicBrainz, rdf-calendar) use RDF to
describe various other kinds of things, such as people, news item
syndication, music, albums, artists, calendars.

Sometimes thesauri cover, to some extent, areas which are also covered
by other RDF-based projects. We are looking for ways to make this
explicit, so that simple-minded machines can query their database and
find matches, regardless of whether a thesaurus-based approach or a
plain-RDF approach was used. 

RDF (and OWL) are at their best when modelling things in terms of
individuals which have properties/attributes and which fall into
classes/categories. Lots of the world can be modelled in terms of things
and categories and properties in this way, which is why RDF appeals. But
RDF's perspective on the world doesn't work so well "out of the box" 
when trying to model things like fluids / mass nouns, or processes. So
there are often enough concepts in a thesaurus that don't have an
obvious correlation in non-SKOS RDF descriptions. That's OK. It's one of
the reasons that SKOS is important. But the concern here is with the
overlap areas, with things (typically individuals/objects, or classes of
individual/object) which are (i) picked out in the SKOS world by being
the basis for 1 or more SKOS concepts (ii) likely also to be described
using various other RDF-based formats.

RDF tries to impose some basic design constraints across all projects
that use it, to make things easier for data-merging, extensibility etc.
What we're doing with skos:represents (or whatever it gets called) is
coming up with a little add-on that helps SKOS-based RDF data work better 
with non-SKOS RDF data. 

For cases like people, places, organisations,
this is quite important, since non-SKOS RDF can directly model some very
detailed characteristics of these entities. We can use non-SKOS RDF to
describe a Person's age, height, employer, etc.; or a place's latitute
and longitude, or an Organization's certifications. That's the sort of
task RDF was built for. But at the same time we can use SKOS RDF to 
describe thesauri which mention those self-same people, places,
organizations, and which treat them explicitly as concepts in a
knowledge organization system. Often these thesauri are used in
bibliographic datasets, library practice etc., and so can be hugely
useful for information retrieval purposes.

The moon on the stick that we're chasing after is to be able to combine
both styles of representation. So for example we might draw upon a
SKOS/RDF thesaurus that included concepts corresponding to various
organisations (eg. UN, CND, Fox Hunters Anonymous etc). And we might
draw upon some non-SKOS RDF to learn more about those organizations, for
eg. that the UN is xyz:headQuarteredIn New York, or was xyz:formed_in
"1942" (or whenever...). 

SKOS is great at hooking into established thesauri, and at representing 
those aspects of the world that aren't so easy to describe in terms of
individuals, classes and properties. Non-RDF SKOS is better than SKOS at
describing arbitrary named properties of individual things, and the class
hierarchies they can be categorised into. Both are useful and valuable,
and the use of RDF means the two can be used alongside each other. What
I'm hoping for is that we can go a little further than simply using
these approaches in parallel, and start to figure out how to build 
little bridges between these complementary styles of representation...

I suspect the reason that this is confusing (apart from my own
limitations at explaining things) is that RDF is simultaneously (i) a
(friendly!) "rival" to the thesaurus-based approach to
describing/classifying/etc (ii) a platform within which thesauri can be
represented pretty much as-is, without having to reorganise them to fit
with RDF's own representational conventions.

In other words, when people ask us "Should I be creating an Ontology or
a Thesaurus?", we want to figure out what to tell them (beyond "yes!")
about the diffent approaches possible using SKOS, RDF, OWL etc. And
hopefully have some conventions for connecting up information created in
either style.


Received on Wednesday, 29 September 2004 18:41:24 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 2 March 2016 13:32:04 UTC