- From: Jakob Voss <jakob.voss@gbv.de>
- Date: Wed, 12 Jul 2006 11:45:37 +0200
- To: public-esw-thes@w3.org
Hi, SKOS seems very useful but after reading the documentation I have three questions about essential properties of a langauge to describe KOS like promised in the abstract of SKOS Core: 1. How to encode Notations This question has been raised in February but a clear answer is missing. Here is yet another possible solution: A notation is a prefered label of an artificial language so just use: <skos:prefLabel xml:lang="art">...</skos:prefLabel> "art" is included in ISO 639-2. Advantage: No change needed Disadvantage: every KOS can only have one notation system 2. How to get the size of a collection When browsing through a KOS you'd like to get the number of items that are indexed with a concept (or with a concept and its subconcepts!). For example see http://dmoz.org/Arts/Costumes/ In this case the number of items in class Arts:Costumes (33) is shown, but there are only 5 items that are directly indexed with Arts:Costumes so you need a way to encode two numbers with a concept - any suggestions who to do it in a consistent way 3. How to encode concept combinations This is slightly complex but important. Maybe it can easily can be solved with SKOS Mapping which I have not looked at enough. In some thesauri you have non-prefered-terms and relations that point you to use a concept in combination with another concept or vice versa to indicate postcoordination or precoordination in indexing. For instance the concept "Spanish town" could be defined as non-prefered and point you to use the two concepts "Spain" and "town" (post-coordination) or if you want to have "Spanish town" as a concept for indexing (pre-coordination) then "Spain" and "town" both have a special kind of associative relationship to "Spanish town". How should I encode this in SKOS? Thanks and greetings, Jakob
Received on Wednesday, 12 July 2006 09:45:10 UTC