W3C home > Mailing lists > Public > public-swd-wg@w3.org > March 2008

Re: ISSUE 77 and postcoordination [and ISSUE-40!]

From: Jakob Voss <jakob.voss@gbv.de>
Date: Fri, 14 Mar 2008 17:06:12 +0100
Message-ID: <47DAA274.3060600@gbv.de>
Cc: public-esw-thes@w3.org, public-swd-wg@w3.org

Hi Stella, Hi Leonard,

Stella wrote:

> Leonard Will wrote:
>> Pre- and post- therefore refer to whether concepts are coordinated 
>> before or after the cataloguing of the resource is completed and 
>> stored ready for retrieval.
 >
> Agreed. And I am therefore puzzled as to why anyone wants to represent 
> postcoordination in SKOS. Postcoordination is something we use at the 
> time of searching, i.e. postcoordination is applied in the queries of 
> searchers, to detect combinations found in the metadata of indexed 
> resources. 

No, postcoordination is also found when creating a vocabulary and when 
indexing resources.

> But it is not part of the vocabulary - any preferred term/concept in
 > a thesaurus can in principle be postcoordinated with any  other
 > preferred term/concept.

Yes, any concepts can be postcoordinated - they don't have to come from 
the same vocabulary. As every concept has an URI, that's no problem.

> If SKOS is for publishing/exchanging vocabularies, it does not need 
 > to do anything about postcoordination.

No, as soon as SKOS is used for more then indexing one resource with one 
concept, we need a way to express postcoordination.

>> As I understand it, the SKOS use of the term "collection" is for 
>> something different altogether. SKOS usage seems to be that 
>> "collection" is what in thesaurus practice we would call an "array", 

Beeing for what in thesaurus practise is called an "array" does not 
hinder to be used for more than this - as long as the semantics don't 
conflict which is not the case in my point of view.

>> i.e. a set of sibling concepts, sharing a common broader parent. This 
>> is part of the structure of a controlled vocabulary, and is not 
>> related to the assignment of more than one term when indexing a resource.
> 
> Agreed.

Well, maybe skos:Concept fits better. Indexing resource <R> with 
concepts <A> and <B> could then be expressed as

<R> skos:subject [ a skos:Concept ; skos:narrower <A>, <B> ] .

This means: <R> is indexed with an abstract, unknown concepts that has 
<A> and <B> as narrower concepts.


I wrote:

>  In particular I found two related gaps in the current draft. First is
> how to encode postcoordination of concepts and second is how to map to
> coordinated concepts. Let me give an example:
 >
>  Given one Concept Scheme with two concepts labeled "holdiay" and 
> "2008":
 >
>  x:holiday a skos:Concept; skos:prefLabel "Holiday" .
>  x:y2k8 a skos:Concept; skos:prefLabel "2008" .
 >
 >  How do you encode that fact that a resource '#R' was indexed with both
> together in a specific context (person, date, etc.)? You somehow 
> have to connect two statements:

Stella wrote:

> Why do this in SKOS? Why not leave it to a downstream application?

Because the only application of SKOS is indexing resources with 
concepts. You can assign labels and relations between concepts but as 
long as you don't use the concepts for indexing stuff, they are pretty 
useless. If there is no easy way to express the basic information of 
indexing (WHO indexed WHAT with WHICH concept(s)) SKOS would be little 
more then a a toy for information scientists. The real action takes 
place in social tagging applications.

I am not sure whether skos:Collection or skos:Concept is the right class 
or whether we must add another class but I am sure that this is the way 
to encode tagging information is

<WHAT> skos:subject [
   rdfs:type <C-X> ;
   <r-X> <WHICH-1> ;
   <r-X> <WHICH-1> ; ...
] .

And it's the same for encoding most of the mappings - which are not just 
1-to-1:

<CONCEPT-1> skos:exactMatch [
   rdfs:type <C-X> ;
   <r-X> <CONCEPT-2> ;
   <r-X> <CONCEPT-3> ; ...
] .

I proposed to use skos:Collection as <C-X> and skos:member as <r-X>.

Greetings,
Jakob

-- 
Jakob Vo▀ <jakob.voss@gbv.de>, skype: nichtich
Verbundzentrale des GBV (VZG) / Common Library Network
Platz der Goettinger Sieben 1, 37073 G÷ttingen, Germany
+49 (0)551 39-10242, http://www.gbv.de
Received on Friday, 14 March 2008 16:07:10 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 14 March 2008 16:07:11 GMT