Re: Why RDF? was: ISSUE-160: Allowing collections in semantic relationships

On Mon, 22 Dec 2008 at 11:15:15, Dan Brickley <danbri@danbri.org> wrote

>On 16/12/08 17:57, Ed Summers wrote:
>> On Tue, Dec 16, 2008 at 10:51 AM, Leonard Will
>> <L.Will@willpowerinfo.co.uk>  wrote:
>>> But if it is just for publication and exchange, why bother putting it into
>>> an RDF framework? There are various simpler formats that have been used in
>>> the past for the exchange of thesaurus data. I thought that the point of RDF
>>> was to make it amenable to machine reasoning.
>>
>> From my perspective the really powerful thing that SKOS provides as a 
>>semantic web application are web identifiers (URIs) for concepts.
>>
>> Minting URIs for your concept schemes and concepts allows them to be
>> referenced and resolved easily. This allows descriptions of resources
>> to unambiguously use SKOS concepts, while also providing a means for
>> mapping concepts from different concept schemes together.

>I found a nice citation for this idea of named concepts in a linked 
>information system:
>
>
>       In http://www.w3.org/History/1989/proposal.html
>       Information Management: A Proposal
>       Tim Berners-Lee, CERN
>       March 1989, May 1990
>
>[[
>The problem with keywords
>
>Keywords are a common method of accessing data for which one does not 
>have the exact coordinates. The usual problem with keywords, however, 
>is that two people never chose the same keywords. The keywords then 
>become useful only to people who already know the application well.
>
>Practical keyword systems (such as that of VAX/NOTES for example) 
>require keywords to be registered. This is already a step in the right 
>direction. A linked system takes this to the next logical step. 
>Keywords can be nodes which stand for a concept. A keyword node is then 
>no different from any other node. One can link documents, etc., to 
>keywords. One can then find keywords by finding any node to which they 
>are related. In this way, documents on similar topics are indirectly 
>linked, through their key concepts. A keyword search then becomes a 
>search starting from a small number of named nodes, and finding nodes 
>which are close to all of them.
>
>It was for these reasons that I first made a small linked information 
>system, not realising that a term had already been coined for the idea: 
>"hypertext".
>]]
>
>cheers,
>
>Dan

This is a good argument for the use of a controlled vocabulary when 
indexing resources. If everyone used the same vocabulary, in which 
concepts are clearly defined and there is agreement on the term to be 
used as the preferred label for each concept, then indeed the 
effectiveness of a retrieval system could be much improved.

In most practical retrieval systems, a single controlled vocabulary is 
used, and it is only necessary to declare what this is, and to make it 
available to indexers and searchers. If URIs are used, this means that 
all the terms used will have the same prefix identifying the scheme.

What seems much more difficult is the apparent expectation that concepts 
from many different schemes can be used at the same time to index 
resources. If they don't belong to the same coherent scheme, the 
relationships between concepts cannot be clearly defined and specified. 
The definition of "poverty", say, in one scheme, may be quite 
inconsistent with the definition of "wealth" in another. The great 
majority of terms are potentially ambiguous: consider "ships, shoes, 
sealing-wax, cabbages and kings". Is "Brussels sprouts" in one scheme a 
narrower concept than "cabbages" in another scheme?

I realise that SKOS provides for mapping between schemes, but any 
mapping cannot have the consistency of a single scheme. The mapping 
relationships provided at present, "broader", "narrower", "exact" and 
"related" do not allow the definitions of concepts in each scheme to be 
adjusted to make them accurate, as can be done within a single scheme, 
and the "overlapping" relationship is missing.

I am concerned that there may be a lack of clarity in people's minds 
between the use of SKOS for encoding a single controlled vocabulary, for 
which it seems more complicated than necessary, and for encoding 
mappings between vocabularies, which is much more difficult. We have not 
addressed mapping yet in the ISO 25964 model, but will have to look at 
it next year.

If you just use concepts from multiple schemes without building 
relationships between then, then you lose the benefits of a structured, 
controlled vocabulary. Certainly you can do this, but users have to be 
aware of the implications.

Leonard

-- 
Willpower Information       (Partners: Dr Leonard D Will, Sheena E Will)
Information Management Consultants              Tel: +44 (0)20 8372 0092
27 Calshot Way, Enfield, Middlesex EN2 7BQ, UK. Fax: +44 (0)870 051 7276
L.Will@Willpowerinfo.co.uk               Sheena.Will@Willpowerinfo.co.uk
---------------- <URL:http://www.willpowerinfo.co.uk/> -----------------

Received on Monday, 22 December 2008 11:54:22 UTC