- From: Rob Atkinson <rob@metalinkage.com.au>
- Date: Wed, 14 Mar 2018 23:31:20 +0000
- To: Armin Haller via GitHub <sysbot+gh@w3.org>
- Cc: public-dxwg-wg@w3.org
- Message-ID: <CACfF9Lx7Oco_MfeR+VvLQrwgGymzra2QsfbYOoHO4JH7zVNaFg@mail.gmail.com>
So regarding inheritance in general - we need to pay attention to entailments and transitivity and test that our models entail correctly. We may need domain-specific entailment rules, or conventions (such as skos:broader used only for direct parent relationships, as a sub-property of skos:broaderTransitive, and skos:broaderTransitive should ne be asserted, but entailed). These rules may need to be captured in documents - but some may also be captured in formalisms such as SHACL. Profiles will also have inheritance entailments as we will need "flattened" profiles that also record the source of inherited constraints. These will need to include the most restrictive constraint - may or may not need to include all the less restrictive constraints. The profiles model i posted could be augmented with SHACL or something to capture these entailment rules for profiles, but where would we attach such rules to DCAT (or leave it like SKOS where such rules are in the supporting documentation) Rob On Thu, 15 Mar 2018 at 08:09 Armin Haller via GitHub <sysbot+gh@w3.org> wrote: > If we want inheritance, Dataset would need to become a subclass of > Distribution. If global domain restrictions are removed, that may actually > be a possibility, but maybe not desirable. > > -- > GitHub Notification of comment by arminhaller > Please view or discuss this issue at > https://github.com/w3c/dxwg/issues/104#issuecomment-373175694 using your > GitHub account > >
Received on Wednesday, 14 March 2018 23:32:26 UTC