danbri: See also DMozRdf in wiki for tools to clean this data.
libby: main SKOS site - with documentation, reports etc
DanC: SKOS-Core 1.0 Guide seems to explain which end is up, along with nice figures and lots of details
DanC: SKOS-Core 1.0 Guide seems to explain which end is up, along with nice figures and lots of details
ear1grey: in summary: is it fair to say that when a term is stable, it is not necessarily complete? is there (or should there be) a set of criteria that should be met so that schema completeness is a consistent concept? and does there exist (or is there a need for) a property which describes the completeness of a vocabulary term?
chaalsNCE: You can define "completeness" so the answer is yes or no. Are there things terms should have but may not have defined (range, label, a label in the right 20 languages to be useful in the EU, or the hundreds spoken in the southern hemisphere)?
chaalsNCE: Since anyone can say anything about anything, do you mean that a schema has a single document that defines it? What about a document derived from triples harvested in the wild - does that count? What if they conflict, or if they use something like Connolly's rangeIntersects property but don't describe a complete range and domain. Or if the range and domain they describe are not complete themselves (that should be easy at least :-)
chaalsNCE: /me would only ride a Unicorn in a hunt for schema completeness :-) But then, define it as having a handful of useful things defined, and it's easy to pin down...
ear1grey: i'm only thinking in terms of simple schema definitions in simple files, rather than harvested content . however, the harvested triples argument is interesting - if a schema is defined by a set of triples that have been harvested (and some additional provenance information - which could be digitally signed in such a way that the provenance is immutable) , then i think the actual provenance wouldn't matter so much as the fact
chaalsNCE: You can define "completeness" so the answer is yes or no. Are there things terms should have but may not have defined (range, label, a label in the right 20 languages to be useful in the EU, or the hundreds spoken in the southern hemisphere)?
chaalsNCE: Since anyone can say anything about anything, do you mean that a schema has a single document that defines it? What about a document derived from triples harvested in the wild - does that count? What if they conflict, or if they use something like Connolly's rangeIntersects property but don't describe a complete range and domain. Or if the range and domain they describe are not complete themselves (that should be easy at least :-)
chaalsNCE: /me would only ride a Unicorn in a hunt for schema completeness :-) But then, define it as having a handful of useful things defined, and it's easy to pin down...
ear1grey: i'm only thinking in terms of simple schema definitions in simple files, rather than harvested content . however, the harvested triples argument is interesting - if a schema is defined by a set of triples that have been harvested (and some additional provenance information - which could be digitally signed in such a way that the provenance is immutable) , then i think the actual provenance wouldn't matter so much as the fact
crschmidt: Already been chumped in the past, but it's a semantic web search