 |
 |
> From: Bernard Vatant <[log in to unmask]> >>> I proposed lately a 5-star scale also >>> for vocabularies [2] >> >> I'd be happy to learn whether PanLex (http://panlex.org) is an example of >> a vocabulary eligible for that rating > We must all be very cautious here on terminology, but the targets of the > 5-star scale for vocabularies are those quite compact formal vocabularies > defining classes (aka types) and properties (aka attributes) expressed in > either RDFS or OWL, with sometimes a pinch of SKOS. Thank you, Bernard. That helps circumscribe the population of potentially 5-star things. > PanLex (which I did not know - thanks for the link) looks to me more like a > lexical data base (dataset) than a vocabulary in the above sense Yes, it's a database of assertians about lexemes in language varieties. It has made use of data from formal vocabularies, but I doubt that it can reasonably be understood as a formal vocabulary itself. It formalizes assertions about lexemes (such as an assertion that one lexeme is a translation of another lexeme). Sets of such assertions can be reported in arbitrary output formats. An overview of the schema is at http://panlex.org/tech/panlex-db-design.pdf, and perhaps this contains an implicit compact formal vocabulary. > See > http://labs.mondeca.com/dataset/lov/details/vocabulary_lvont.html Both the lvont ontology and GOLD seem capable of representing many of the ideas implicit in the PanLex schema. > from a legal viewpoint, > it's a MUST : you have to understand equivalent terms the same way. No > choice :) Yes, so they can disagree on what that identical understanding must be, choosing whichever equivalent helps prove their claims. > would often define skos:closeMatch relationships Doesn't the definition of this property (including nontransitivity) resemble what most translations in ordinary plurilingual dictionaries assert?