RE: merging and mapping

This conversation is making me nervous about the misunderstandings that
arise when we use our own terminology imprecisely.

By "mapping" I mean setting up a relationship between terms or codes
that represent concepts in different vocabularies. 
By "merging" I mean taking two or more vocabularies and making them into
one vocabulary. (For this purpose I would make use of any mappings that
were available).

But the most obvious applications of mappings do not involve merging.
For example, you can use mappings to translate metadata descriptions
from one vocabulary to another. Or you can use them to translate Search
statements from one vocabulary to another. You can do either of these
without building a merged vocabulary. 

Any of the applications will work much better if the people who built
the vocabularies and the mappings followed the guidelines (e.g. ISO
2788) rigorously while doing it. 

Between 2 different vocabularies, you can and should expect to find one
label applying to 2 different concepts. But within one thesaurus, it is
straightforward to require each term to be unique, and not an
unreasonable requirement.

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 Charles
McCathieNevile
Sent: 15 March 2004 14:17
To: Miles, AJ (Alistair) 
Cc: public-esw-thes@w3.org
Subject: Re: merging and mapping



Clearly Leonard and I (and to some extent Alistair and I) are currently
looking from different perspectives. I don't see any point in a mapping
unless it is to allow use of two vocabularies as if they are merged (an
artifact of RDF is that as soon as we have a mapping we can treat these
things as merged).

As I understand it one of the important features of the work is the
ability to do this mapping (although it isn't kept in the same
namespace, it is a requirement, and it is certainly the basis of my
interest in this).

I maintain that one of the use cases that makes the SKOS work
interesting is the parallel work on mapping, and that we should ensure
that we don' break this.

I therefore think that we should separate what we consider Best Practice
(for example having the same label as possibly referring to two concepts
is generally a bad idea) from making something absolutely incorrect when
there is reason to believe that it will arise in practice due to
sensible work. So I would continue to suppport the idea that things are
"strong usage recommendations" (but not encoded in the ontology
definition frameworks), rather than seeking to outlaw them by specifying
appropriate RDF/OWL constraints.

Cheers

Chaals

On Mon, 15 Mar 2004, Miles, AJ (Alistair)  wrote:

>> Chaals raises the issue of merging different vocabularies. I
[snip]
>> above applies there too.  As the present SKOS draft doesn't cover 
>> questions of mapping between vocabularies, I think we should keep it 
>> to what is necessary and desirable for a single vocabulary.
>>
> I think 'merging' and 'mapping' are completely different scenarios.  
>In mapping, the schemes are kept separate, and linked via mapping 
>statements. In merging, a new scheme is created by combining concepts 
>from different sources.
>
>It's worth considering the situations where separate groups of people 
>(part of a larger community) are responsible for contributing concepts 
>as part of a larger scheme.  In this scenario we may expect to find 
>overlap of labels.

Received on Monday, 15 March 2004 10:13:10 UTC