search

w3c / dxwg

Data Catalog Vocabulary (DCAT)
https://w3c.github.io/dxwg/dcat/
Other
139 stars 55 forks source link

Are inference rules constraints that must be followed by profiles? #294

Open azaroth42 opened 6 years ago

azaroth42 commented 6 years ago

For example, PROV-O has a long document of inference rules, including one that we ran into in the Annotation WG when trying to use generatedBy: https://www.w3.org/TR/prov-constraints/#generation-generation-ordering_text

The side effects of this rule on an otherwise very useful term meant that we felt it was not usable by us, and we had to look for a different term. Should profiles also be required to recognize inference rules?

I propose that they should, based on the discussion on the call 2018-07-03 where we discussed a definition of profile that it must only further constrain and not expand the definition of the terms that it includes. The implication is that already very constrained vocabularies and models, such as PROV-O, are significantly less likely to be used in profiles for fear of running into unexpected inferences.

rob-metalinkage commented 6 years ago

An interesting question - because different profiling mechanisms have different levels of expressivity - so in general if you rely on OWL in your profile I dont think you can override semantics defined by OWL. constructs ..

For now, I think we should make sure we can cover the typical case (conforming constraints) , until such time as we have a detailed Use Case and examples for why something that does not conform to a base specification should be treated as a profile (perhaps a profile is a special case of a more general thing that has weaker requirements?)