W3C home > Mailing lists > Public > public-esw-thes@w3.org > August 2006

RE: SKOS Mapping Vocabulary Specification query

From: George Macgregor <george.macgregor@strath.ac.uk>
Date: Thu, 17 Aug 2006 10:27:39 +0100
To: "'Tudhope D S \(AT\)'" <dstudhope@glam.ac.uk>
Cc: <public-esw-thes@w3.org>
Message-ID: <E1GDe9n-0002Bb-VL@esna.cc.strath.ac.uk>

Hi Douglas,

> I wonder if slightly different situations are being discussed here?

Yes - and it's my fault!  My original message was simply to determine the
status of the SKOS MVS, whether any extensions are being proposed and their
potential nature.   

The NKOS Workshop is definitely the best place to discuss these issues
further as numerous interesting sub-issues emerge. Unfortunately I am unable
to attend.  My colleague Dennis Nicholson will be there to present his paper
and I am sure he will contribute meaningfully to all the workshop
discussions.  

Thank you to everyone for your comments.

George
----------------------------------------------
George Macgregor,
Centre for Digital Library Research (CDLR),
Department of Computer & Information Sciences,
University of Strathclyde, Livingstone Tower,
26 Richmond Street, Glasgow, UK, G1 1XH
tel: +44 (0)141 548 4752
fax: +44 (0)141 548 4523
web: http://cdlr.strath.ac.uk/
-------------------------------------------- 
________________________________________
From: Tudhope D S (AT) [mailto:dstudhope@glam.ac.uk] 
Sent: 16 August 2006 20:32
To: public-esw-thes-request@w3.org
Cc: george.macgregor@strath.ac.uk
Subject: RE: SKOS Mapping Vocabulary Specification query

I wonder if slightly different situations are being discussed here? 
 
The SKOS MVS and the BSI standards are primarily intended to express a
mapping between two KOS (vocabularies). They operate at the concept level.
As part of the process in creating a mapping, a decision may have been made
between various candidate term matches (at finer levels of detail). 
 
HILT is concerned (1) with mapping different KOS to the DDC and (2) with
resolving (mapping) a user query to controlled terminology (DDC). The latter
case would involve the finer level of detail (spelling variations etc) you
describe. 
 
As Stella describes, a concern with free text or NLP issues may also involve
finer levels of detail.
 
The NKOS Workshop at ECDL would be  agood place to further discuss these
issues since there are presentations by both HILT and the FAO on mapping -
http://www.ukoln.ac.uk/nkos/nkos2006/
 
Doug
 
 
Douglas Tudhope 
Reader, School of Computing 
University of Glamorgan 
Pontypridd CF37 1DL 
Wales, UK 
Tel +44 (0) 1443-482271 
Fax +44 (0) 1443-482715 
dstudhope@glam.ac.uk 
http://www.comp.glam.ac.uk/pages/staff/dstudhope 
Editor : The New Review of Hypermedia and Multimedia 
 
 
>>>>>>>>>>>>>>>>
HILT (http://hilt.cdlr.strath.ac.uk/) also assumes that items have been
indexed by humans using controlled vocabularies.  HILT III is supposed to
function in an M2M environment and is supposed to assist users at the client
end (predominantly HE and FE students, academics, researchers) to
simultaneously interrogate a variety of distributed services using different
vocabularies - and therefore using various terms to describe the same
concepts. 

At the moment HILT III is experimenting with LCSH, UNESCO, HASSET, AAT,
MeSH, NMR, and a number of other terminologies.  These terminologies are
popular within the UK JISC Integrated Information Environment and are used
by many services, repositories and digital collections.  Each of these
terminologies is mapped (by HILT) to DDC which functions as a central spine
(similar to a switching language) which the user then uses to disambiguate
queries across multiple terminologies.  Details of any mapped terms matching
a user query - complete with match type(s) - are then sent by HILT to the
client via SOAP using SKOS-Core and (hopefully) the SKOS Mapping Vocabulary
Specification. 

Thanks for your comments.  We will keep you posted on the progress!

Regards,

George
----------------------------------------------
George Macgregor,
Centre for Digital Library Research (CDLR),
Department of Computer & Information Sciences,
University of Strathclyde, Livingstone Tower,
26 Richmond Street, Glasgow, UK, G1 1XH
tel: +44 (0)141 548 4752
fax: +44 (0)141 548 4523
web: http://cdlr.strath.ac.uk/
--------------------------------------------

> -----Original Message-----
> From: public-esw-thes-request@w3.org [mailto:public-esw-thes-
> request@w3.org] On Behalf Of Stella Dextre Clarke
> Sent: 16 August 2006 15:03
> To: 'George Macgregor'; 'G.hopmans'
> Cc: public-esw-thes@w3.org
> Subject: RE: SKOS Mapping Vocabulary Specification query
>
>
> George,
> Gosh, it never occurred to me that people would want to distinguish
> singular/plural etc. But then with BS 8723 ( and to some extent with
> SKOS too) we are coming from an environment in which we assume that
> documents (items) have been indexed by human indexers using a controlled
> vocabulary so that they can identify which are the important concepts in
> the document and call each such concept by one consistent name. In other
> words, singulars, plurals and other variants are deliberately brought
> together in one name, *so long as they have the same meaning*.
>
> However, if you were thinking of vocabularies to help with free text,
> then you might well want to distinguish singulars and plurals. Or you
> might be a linguist looking at patterns of term usage, and want to
> analyse frequency of occurrence of the different usages.
>
> If folksonomies are included among the vocabularies, that's different
> again.
>
> So before deciding how to do the mapping, it would be useful to have a
> clear vision of who is the user, what are they mostly trying to do, have
> the searchable resources been indexed with a controlled vocabulary, and
> are the users trying to retrieve concepts or just words. Also, do the
> users understand what a controlled vocabulary does for them.
>
> Have fun
> 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: George Macgregor [mailto:george.macgregor@strath.ac.uk]
> Sent: 16 August 2006 12:53
> To: 'Stella Dextre Clarke'; 'G.hopmans'
> Cc: public-esw-thes@w3.org
> Subject: RE: SKOS Mapping Vocabulary Specification query
>
>
> Hi Stella, Gabriel
>
> Thanks for your emails - much appreciated.
>
> > Well, this may not be exactly what you are asking about, but we hope
> > it will be useful for a lot of applications.  Let me know if you'd
> > like to see a draft.
>
> This sounds very useful indeed and it will definitely inform our
> approach. Thanks!  We will keep everyone posted on developments!
>
> The big question for HILT relates to the type of approach.  Recent
> approaches adopt a conceptual appraisal of matching, and this is
> certainly consistent with how SKOS functions and current thinking
> generally (such as the recent Renardus project).  Yet, a large
> proportion of mapping work in the past considered this inadequate and
> adopted an approach of distinguishing matches at a finer level of detail
> (e.g. singular/plural matches, spelling variations, etc., in addition to
> concept matches and exact matches).  Similarly, previous work in HILT
> (much of it anecdotal,
> admittedly) appeared to infer that users required such detail to
> understand why they had received particular mappings and to enable them
> to formulate similar queries in the future.  It was also necessary to
> allow users to disaggregate huge result sets which might normally (under
> SKOS for example) be conceptually equivalent but which in reality
> subsumed various types of 'exact match'.  For example, to-the-letter
> exact matches or matches with spelling variations or intervening
> characters, and so forth.
>
> Clearly this area lacks sufficient user research as there are numerous
> assumptions in the literature as to what users might find useful! :-)
> Perhaps this is something we can investigate this during the project
> lifetime...
>
> Thanks to all.
>
> George
> ----------------------------------------------
> George Macgregor,
> Centre for Digital Library Research (CDLR),
> Department of Computer & Information Sciences,
> University of Strathclyde, Livingstone Tower,
> 26 Richmond Street, Glasgow, UK, G1 1XH
> tel: +44 (0)141 548 4752
> fax: +44 (0)141 548 4523
> web: http://cdlr.strath.ac.uk/
> --------------------------------------------
>
> > -----Original Message-----
> > From: Stella Dextre Clarke [mailto:sdclarke@lukehouse.demon.co.uk]
> > Sent: 15 August 2006 20:49
> > To: 'George Macgregor'
> > Cc: public-esw-thes@w3.org
> > Subject: RE: SKOS Mapping Vocabulary Specification query
> >
> > George,
> >
> > You may like to consider the mapping types in the forthcoming BS 8723
> > Part 4. This standard will define mappings as "statements of the
> > relationships between the terms, notations or concepts of one
> > vocabulary and those of another". It will describe the following types
>
> > of mapping:
> > - equivalence (comparable to the equivalence relationship in a
> > multilingual thesaurus)
> > - hierarchical (comparable to the BT/NT relationship in a thesaurus)
> > - associative (comparable to the RT/RT relationship in a thesaurus)
> > The main emphasis in the standard is on equivalence mappings.
> >
> > It will describe degrees of equivalence, acknowledging that some
> > proposed equivalents are partial or inexact (but may nonetheless be
> > acceptable in appropriate contexts) without establishing distinct
> > mapping types for the more and less exactly equivalent respectively.
> > It will also suggest a method of handling one-to-many equivalences,
> > i.e. where a concept in one of the vocabularies is represented by a
> > combination of concepts in the other vocabulary, and for this
> > situation two different combination types are described.
> >
> > Well, this may not be exactly what you are asking about, but we hope
> > it will be useful for a lot of applications.  Let me know if you'd
> > like to see a draft.
> >
> > 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 George Macgregor
> > Sent: 15 August 2006 15:59
> > To: public-esw-thes@w3.org
> > Subject: SKOS Mapping Vocabulary Specification query
> >
> >
> >
> > Dear all,
> >
> > At the Centre for Digital Library Research (CDLR) we are in the
> > process of researching and developing a M2M mapping-based terminology
> > server. This work is being undertaken via the third phase of the
> > High-Level Thesaurus project (HILT) [1].  HILT III will offer web
> > services access (via SOAP) and plans to use SKOS Core for wrapping
> > terminology sets for sending. In particular, this would entail the use
>
> > of the SKOS Mapping Vocabulary Specification.
> >
> > Previous phases of HILT indicated that a degree of specificity was
> > required when characterising mapping match types.  This was necessary
> > to facilitate improved results ranking and to provide users with
> > adequate relevance feedback.  Vocabulary switching work conducted by
> > Chaplan [2] identified 19 separate mapping match types and noted the
> > difficulty in using conceptual approaches to denote equivalence.  More
>
> > recently, Liang et al. [3] have found problems in implementing the
> > SKOS matches for mapping between disparate thesauri.  This was
> > attributed to difficulties pertaining to the match type definitions
> > and their ability to accommodate more complex matches.
> >
> > Clearly there is a need to offer a simple conceptual approach to
> > specifying equivalence - and this is what the SKOS MVS does; however,
> > we would be keen to know the status of the SKOS MVS, whether any
> > extensions are being
> > proposed and, if so, their nature.   Since there has not been
> > significant
> > public comment on the SKOS mapping types, we would also be keen to
> > hear comments and whether people consider the MVS to be sufficiently
> > specific. Our current thoughts are that we might extend the SKOS MVS
> > to incorporate extra match types, but we would want to do this
> > collaboratively and in a way that can accommodate any foreseeable (or
> > potential) additions to the SKOS MVS.
> >
> > Thanks in advance.
> >
> > George
> >
> > References:
> >
> > [1] HILT: http://hilt.cdlr.strath.ac.uk/
> >
> > [2] Margaret A. Chaplan, Mapping Laborline Thesaurus terms to Library
> > of Congress Subject Headings: implications for vocabulary switching,
> > Library Quarterly 56(1) (1995) 39-61.
> >
> > [3] A. Liang, M. Sini, Chang Chun, Li Sijing, Lu Wenlin, He Chunpei
> > and J. Keizer, The mapping schema from Chinese Agricultural Thesaurus
> > to AGROVOC, 6th AOS Workshop - Thesaurus enrichment and facilitating
> > interoperability through mapping (2005).  Available:
> > ftp://ftp.fao.org/docrep/fao/008/af241e/af241e00.pdf
> >
> > ----------------------------------------------
> > George Macgregor,
> > Centre for Digital Library Research (CDLR),
> > Department of Computer & Information Sciences,
> > University of Strathclyde, Livingstone Tower,
> > 26 Richmond Street, Glasgow, UK, G1 1XH
> > tel: +44 (0)141 548 4752
> > fax: +44 (0)141 548 4523
> > web: http://cdlr.strath.ac.uk/
> > --------------------------------------------
> >
> >
> >
>
>
>
Received on Thursday, 17 August 2006 09:27:39 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:38:54 GMT