W3C home > Mailing lists > Public > public-esw-thes@w3.org > March 2007

RE: [SKOS] RE: ISSUE-33: GroupingConstructs

From: Stella Dextre Clarke <sdclarke@lukehouse.demon.co.uk>
Date: Tue, 27 Mar 2007 18:33:38 +0100
To: "'Miles, AJ \(Alistair\)'" <A.J.Miles@rl.ac.uk>, "'SWD WG'" <public-swd-wg@w3.org>
Cc: <public-esw-thes@w3.org>
Message-ID: <002001c77096$113d8c00$0300000a@DELL>

Alistair,

You ask for clarification re RTs, and I wish I could give you a simple
Yes or No. But to be strictly accurate it is a little more complicated
than that...

Firstly, neither ISO 2788 nor BS 8723 currently gives you a straight
answer on this point. To follow BS 8723 strictly, you ought to inspect
each candidate RT relationship (e.g. between "agricultural industries"
and "dairy personnel" to use your example) to see whether it qualifies.
To qualify, the pair should be semantically or conceptually associated
to such an extent that an explicit link would suggest additional or
alternative terms for use in indexing or retrieval. And the judgment is
liable to vary from one context to another.

In practice, it is quite time-consuming to consider each of those pairs
individually. And in most cases, the concepts would not have been placed
in such close juxtaposition in the systematic display if there was not a
pretty strong association between them. On those grounds, it may be
possible to justify automatic creation of RT links in all cases in the
situation you describe.

AS it happens, the only examples of published thesauri I can think of
that more or less follow the guidelines in BS 8723 for node labels and
have linked alphabetical and classified displays in this sort of style,
were developed or guided or at least inspired by a lady called Jean
Aitchison, the great pioneer and guru of thesaurus development in the
UK. Through her influence, over the years two software suites have been
developed to manage her thesauri. Both of these software products manage
RT relationships in the way that you are surmising. In other words, all
the concepts in an array are deemed to have a hierarchical relationship
with the concept above them, unless a node label in between introduces a
new facet; in the latter case, the relationship generated is RT instead
of BT/NT.

OK, that's a little look behind the scenes. For the purposes of SKOS, I
think you have to assume there is no relationship between the concepts
"above" and "below" a node label, unless there is some explicit clue in
addition to the node label.

Stella

*****************************************************
Stella Dextre Clarke
Information Consultant
Luke House, West Hendred, Wantage, Oxon, OX12 8RR, UK
Tel: 01235-833-298
Fax: 01235-863-298
SDClarke@LukeHouse.demon.co.uk
*****************************************************



-----Original Message-----
From: Miles, AJ (Alistair) [mailto:A.J.Miles@rl.ac.uk] 
Sent: 27 March 2007 16:22
To: Stella Dextre Clarke; SWD WG
Cc: public-esw-thes@w3.org
Subject: RE: [SKOS] RE: ISSUE-33: GroupingConstructs


Hi Stella,

Thanks so much for this information.

Can I ask you to clarify a small point...

BS 8723-2:2005 gives the following example (abridged) for the use of
node labels to introduce new facets in a systematic display:

agricultural industries

  (people)
  farm managers
  dairy personnel
  shepherds

  (products)
  cereal products
  dairy products

Would the alphabetical display of the same thesaurus show the following:

agricultural industries
  RT farm managers
  RT dairy personnel
  RT shepherds
  RT cereal products
  RT dairy products

...?

Thanks,

Alistair.

> -----Original Message-----
> From: Stella Dextre Clarke [mailto:sdclarke@lukehouse.demon.co.uk]
> Sent: 27 March 2007 11:35
> To: Miles, AJ (Alistair); 'SWD WG'
> Cc: public-esw-thes@w3.org
> Subject: RE: [SKOS] RE: ISSUE-33: GroupingConstructs
> 
> Alistair,
> I've just been looking at your blog statement at 
> http://www.w3.org/2006/07/SWD/wiki/SkosDesign/GroupingConstructs
> and I think it is a useful and accurate description of the situation 
> we're trying to work through.
> 
> You gave a caution about possible inconsistencies between ISO 2788 and

> BS 8723-2, And I'll now try and cast a little light on that.
> 
> The use of the term "facet indicator" in ISO 2788 was probably a 
> mistake. We deliberated over this while developing BS 8723, and spoke
to
> one of the persons who was probably influential in putting this term 
> into ISO 2788. After careful consideration, she was able to advise us
to
> drop the term "facet indicator" in this context. BS 8723-2 instead 
> speaks of "characteristics of division", in line with current thinking

> on faceted classification (at least as it is practised in the UK!).
> 
> The other small inconsistency you were probably referring to concerns 
> the second type of node label, where characteristics of division are
not
> involved. In this second type, the node labels introduce "different 
> types of concepts" according to ISO 2788, or "new facets" according to

> BS 8723. I don't see any discrepancy here, since a facet is defined as
a
> "grouping of concepts of the same inherent category".  (It is possible

> that the ISO 2788 committee was trying to avoid getting too deeply 
> dependent on the specialist terminology of the faceted classification
> community.)
> 
> So as far as ISO 2788 and BS 8723 are concerned, I believe we have 
> pretty good continuity. But the AAT with its guide terms is definitely

> different.  The introduction to the printed thesaurus states that
guide
> terms "are also called node labels or facet indicators by some 
> thesauri." Well, that may be true of some thesauri, but not of ISO
2788
> or BS 8723! The way the AAT does things was very carefully planned for

> its own context, subtly different from many information retrieval 
> applications and the conventional thesaurus. For example, if you look
at
> the AAT online or in print, you won't find the tags "BT" or "NT" 
> anywhere (although you do find "RT" occasionally in the printed 
> version). So they nowhere fall into the trap of presenting a guide
term
> as an "NT" of one of the descriptors recommended for indexing.
However,
> they say that guide terms "are used as broader terms to some groups of

> descriptors when there is no broader term to bring the grouping term 
> together. ....In addition to serving as broader terms, many guide
terms
> define the characteristics of division by which a listing of narrower 
> terms is clustered."
> 
> You could summarise the above by saying that there is a lot of common 
> ground between AAT guide terms and ISO 2788 node labels. For example, 
> there are two different types of guide terms, and the AAT makes it 
> abundantly clear that the guide terms are not to be used for indexing.

> But the requirement for some guide terms to serve as broader terms 
> (though only for presentation/grouping purposes, not for indexing) is
an
> important difference.
> 
> I'm not sure how you get SKOS to work well with both approaches 
> simultaneously. All the best!
> Stella
> 
> 
> *****************************************************
> Stella Dextre Clarke
> Information Consultant
> Luke House, West Hendred, Wantage, Oxon, OX12 8RR, UK
> Tel: 01235-833-298
> Fax: 01235-863-298
> SDClarke@LukeHouse.demon.co.uk
> *****************************************************
> 
> 
> 
> -----Original Message-----
> From: public-esw-thes-request@w3.org 
> [mailto:public-esw-thes-request@w3.org] On Behalf Of Miles, AJ
> (Alistair)
> Sent: 26 March 2007 13:06
> To: SWD WG
> Cc: public-esw-thes@w3.org
> Subject: [SKOS] RE: ISSUE-33: GroupingConstructs
> 
> 
> 
> Guus asked me to suggest an issue to open at the next telecon. I
> suggest:
> 
> http://www.w3.org/2006/07/SWD/track/issues/33
> 
> Cheers,
> 
> Alistair.
> --
> Alistair Miles
> Research Associate
> CCLRC - Rutherford Appleton Laboratory
> Building R1 Room 1.60
> Fermi Avenue
> Chilton
> Didcot
> Oxfordshire OX11 0QX
> United Kingdom
> Web: http://purl.org/net/aliman
> Email: a.j.miles@rl.ac.uk
> Tel: +44 (0)1235 445440
> > -----Original Message-----
> > From: public-swd-wg-request@w3.org
> [mailto:public-swd-wg-request@w3.org]
> > On Behalf Of SWD Issue Tracker
> > Sent: 26 March 2007 12:58
> > To: public-swd-wg@w3.org
> > Subject: ISSUE-33: GroupingConstructs
> >
> >
> >
> > ISSUE-33: GroupingConstructs
> >
> > http://www.w3.org/2006/07/SWD/track/issues/33
> >
> > Raised by: Alistair Miles
> > On product: SKOS
> >
> > Thesaurus standards describe the use of "node labels" for convenient

> > grouping and display. Many thesauri use node labels in this way.
Some
> > thesauri use "guide
> > terms" for a similar purpose. In some social tagging systems, tags
can
> > be
> > grouped into "bundles".
> >
> > The current vocabulary support in SKOS (called "Collections") for
> these
> > kinds of
> > grouping construct is broken, in that it introduces a contradiction
> into
> > the
> > SKOS specifications.
> >
> > See further information at:
> >
> > http://www.w3.org/2006/07/SWD/wiki/SkosDesign/GroupingConstructs
> >
> >
> 
> 
> 


--
Alistair Miles
Research Associate
CCLRC - Rutherford Appleton Laboratory
Building R1 Room 1.60
Fermi Avenue
Chilton
Didcot
Oxfordshire OX11 0QX
United Kingdom
Web: http://purl.org/net/aliman
Email: a.j.miles@rl.ac.uk
Tel: +44 (0)1235 445440
Received on Tuesday, 27 March 2007 17:33:57 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:38:55 GMT