Re: Comments on SKOS Extension for Sub. Indexing

Antoine Isaac wrote:

>> By the way the properties skos:primarySubject 
>> skos:isPrimarySubjectOf should also be skipped (do you read this
>> Mike? ;-) because they are not well defined (what is "primary"? why
>> isn't there a "secondary"? or "third?" What about weighted index
>> terms? To support syntactic indexing in SKOS first the Coordination
>> issue needs to be solved. And I am not sure whether syntactic
>> indexing is really an SKOS issue. The most important is maybe
>> *ordered sets* of index terms.
> 
>> This reads weird. In my opinion this 'ordered set' idea raises the
>> same problem as the primary or secondary subject one... If you have
>> a motivation for an ordered set, then you have specific indexing
>> needs, an it would perhaps be more adequate to include your own
>> subject specializations (unless you plan unlimited number of
>> ordered subjects)

Ordered index terms (that is to keep in information in which order the
index terms were entered) are fundamental. The difference between "term1
term2, term3" and "term3, term2, term1" might be small but just skipping
this information is an error. SKOS needs a concept of ordered index sets
and this concept is relevant both to coordination for combined concepts
(especially when mapping) and to indexing.

> For coordination it is still in the air, cf the list of issue (and
> requirement) that are being currently worked out on the Semantic Web
> Deployment wiki [1] For notations, well we could raise a dedicated
> issue, but it would be nice for SWD to have more motivating use cases
> than the single Iconclass one I contributed. I suppose UDC would be
> enough, but have no time to write it down now :-(

Almost *any* classification system uses notations and titles instead of
labels only. Because SKOS was designed for thesauri the problem was not
visible before. But in a classification there are two kinds of labels:

1. notations: mostly artificial codes that uniquely identify a concept
2. titles: human readable label, does not need to be unique. May occur
in multiple languages and variants.

This is *not* the same distinction as skos:prefLabel and skos:altLabel.
For use cases see:

http://en.wikipedia.org/wiki/Category:Classification_systems

For example:

http://www.ams.org/msc/
http://publish.aps.org/PACS/
http://www.wipo.int/classifications/ipc/ipc8/
..

Greetings,
Jakob

Received on Monday, 5 February 2007 12:53:58 UTC