Aligning an Ontolex resource in EDOAL
Roland Wingerter
- OWL: Class, Property
- SKOS: Concept, Scheme
- Ontolex: Lexicon, Lex.Entry
Armando Stellato
Thanks Roland for spotting this!
We will definitely add that.
Kind Regards,
Armando
--
You received this message because you are subscribed to the Google Groups "vocbench-user" group.
To unsubscribe from this group and stop receiving emails from it, send an email to
vocbench-use...@googlegroups.com.
To view this discussion on the web visit
https://groups.google.com/d/msgid/vocbench-user/3f5b29f6-37be-4fde-8085-e5bdca511f11n%40googlegroups.com.
Roland Wingerter
Roland Wingerter
Armando Stellato
Dear Roland,
glad you appreciated :-) btw, thanks once more for your feedback on the EDOAL project, the tabs organizer (allowing you to select which tabs you want to see) has been added for Ontolex projects shown in EDOAL projects and the concept tab has been activated by default.
As a side note (it’s not related to VB), a couple of comments about the relationships you found and the one you established:
- The one you found is underspecified: skos:mappingRelation should be an abstract property, an umbrella over the skos mapping properties that should be used
- The one you set is owl:sameAs, however for SKOS resources this is discouraged. The rationale for this lies in the strong implications of owl:sameAs which is not simply suggesting that those two concepts mean the same thing, rather they are exactly two names for the same logical object. While this may sound as the same thing (in a Tarskian perspective, if the interpretation on the domain is the same because “they mean the same thing”, then those logical objects are simply two referents for the same domain object), in the case of owl:sameAs this is explicit and a reasoner treats them as the same logical object, thus you might even incur in inconsistencies between the intensions of two linked thesauri (which, we know, are usually large and not characterized by a rigorous logical modeling), while by using some skos mapping property a reasoner would treat them as different objects (because the semantics of skos mapping relations do not imply any formal connection between the linked nodes), thus the mapping relations are only intended for applications which, looking at them, may define proper navigation/linking across the thesauri. For instance, in your case you could use a skos:exactMatch relation, and an application performing, say, semantically indexed information retrieval could respond to a query performed on your thesaurus, by retrieving documents that are indexed with the dbpedia “radius bone” concept because of their exactMatch connection.
Kind Regards,
Armando
To view this discussion on the web visit https://groups.google.com/d/msgid/vocbench-user/2d478927-c588-4dae-b12c-ea12c0e13481n%40googlegroups.com.
Roland Wingerter
In this case an underspecified relationship
(skos:mappingRelation) is good enough for me, as there is nothing else to start from.
My validation process depends
on the ability to distinguish the original mappings from my own. This makes it possible to report exactly
what has been changed (triples added/deleted). And it requires only a minimum of manual input.
Kind regards
Roland
Roland Wingerter
Roland Wingerter
Armando Stellato
Oh, ya, that’s very useful in Ontolex, especially for large lexical databases..
We should advertise more certain features. However, for those that come handy for performance reasons, we have put alerts in VB. Whenever something to show is too cumbersome, VB alerts the user that another view might be a better choice for the specific dataset,
Cheers,
Armando
To view this discussion on the web visit https://groups.google.com/d/msgid/vocbench-user/c1ab531d-4ba8-4e69-acdc-e669bae81ab0n%40googlegroups.com.