Re: [dxwg] Rename Resource Descriptor class (#573)

Thanks for weighing in - I tjhink @aisaac characterisation is correct - the model is OK but we want to choose the best possible names for things.. 

There is no problem using DCAT for an alternative way of specifying the relation (its an open world assumption in play) - if you do that it might entail under OWL reasoning that a Profile is a dcat:Resource (actually a dcat:Dataset if the resource is a dcat:Distribution which is why formal alignment felt too hard - too much potential argument about what a "Dataset" is !)

So, i think its fine if an implementer chooses to use polymorphism to map the Profile pattern onto the identical DCAT pattern - and an the idea is that an informative alignment is published to do this (again following DCAT precedent exactly).

I dont mind "ProfileResource" at all - we were playing with "ProfileDistribution" - but maybe that makes it too hard to characterise things like guidance notes and mappings to other related models - which may be critical to understand the semantics of a profile. Such things are really not "Distributions" of the Profile at all - they are resources that play a role.

So with all the discussion it seems there are three valid choices for the name of the association class :
1) prof:ResourceDescriptor
2) prof:ProfileResource
3) prof:Resource

(the latter taking Simon's point that Profile is redundant if scoped to prof: and we just need to get over thinking the same local name has any significance outside the namespace scope!)

I can live with any of these - but not something with a narrower implication about the type of resource or its specific role, as we have mechanisms to handle those aspects in a better way.





-- 
GitHub Notification of comment by rob-metalinkage
Please view or discuss this issue at https://github.com/w3c/dxwg/issues/573#issuecomment-532002344 using your GitHub account

Received on Monday, 16 September 2019 23:58:11 UTC