RE: Profile definitions / DCAT update

Not sure about ‘DCAT Core’.
I see three types of guidance:


1.       DCAT vocabulary – extensions and revisions to the gamut of classes and properties in the DCAT namespace, and also axioms that link classes and properties from other vocabularies into DCAT (i.e. domain/range axioms, OWL constraints)

2.       DCAT best practices and usage guidelines – this would include recommendations about the use of non-DCAT classes and properties that might satisfy some of the requirements, while not being axiomatically bound into the DCAT vocabulary. These would be tied to specified use-cases, and would overlap with …

3.       DCAT profiles – formally defined using SHACL/ShEX where possible, possibly (also) as an Ontology graph that <imports> the dependencies if they can be expressed as RDFS/OWL axioms.

From: Rob Atkinson [mailto:rob@metalinkage.com.au]
Sent: Friday, 5 January, 2018 11:32
To: Rob Atkinson <rob@metalinkage.com.au>
Cc: kcoyle@kcoyle.net; public-dxwg-wg@w3.org
Subject: Re: Profile definitions / DCAT update


Re lists and deliverables...

This is actually not a trivial issue - for each requirement there are multiple possible ways to deliver - i.e. should it be promoted to an extension of DCAT or is there an existing vocabulary we should re-use and hence it should be in the guidelines.  In fact, there are probably formalisms of DCAT profiles we can/should deliver, at the very least to support with exemplars and experience the guidelines for how to profiles.

So I suggest the triage might be:

Look at each requirement and decide if a case can be made for all the extra work of a DCAT extension,  and if not, can a top-level general profile be identified - i.e. should we propose a DCAT profile for scientific data (and then delegate it to either a keen subgroup to add as a deliverable, or describe the potential scope in the guidleline deliverable and then park the issue)?

Only generally applicable cases should make it into DCAT core. Maybe we should create a register of possible abstract profiles, (which would be refined for specific application domains) where external vocabularies we decide we want to exploit are introduced - e.g. DCAT -PROV .  Note that profiles describe conformant subsets, and are thus not disjoint, so something can be both DCAT-Prov and DCAT-EU, so this is a matter of high level functional requirements, not a binning/classification exercise.

Rob





On Fri, 5 Jan 2018 at 11:21 Rob Atkinson <rob@metalinkage.com.au<mailto:rob@metalinkage.com.au>> wrote:

Re profiles discussion..

As part of the work I am doing with the OGC developing a Linked Data view of a knowledge base of specifications, profiles and supporting artefacts I find the need to have a formal ontology to define profiles, and the nature of artefacts that support validation and implementation of these profiles (like SHACL documents, schematron, human readable documents, testing guidelines, unit tests etc. I have also been discussing activities with the Australian Government Linked Data working group, where a key concern is how profiles of ISO standards relate to Australian Standards and industry de jure and de facto standards, and the general desire to have formal ontologies to back up descriptions of how such things are linked.

so, as part of the process of "define what we mean by profiles" I propose to submit a straw man OWL model and try to keep it in sync with the semantics of any text definitions, and propose this as an extra deliverable, which with OGC and at least one other we should be able to meet requirements for a rec track for this.

Have put the OGC view (based on ISO concepts)

Rob Atkinson.



On Fri, 5 Jan 2018 at 09:30 Karen Coyle <kcoyle@kcoyle.net<mailto:kcoyle@kcoyle.net>> wrote:
As we all come out of our holiday haze, it is time to get ready for some
important work on DCAT 1.1 and application profiles. Here are two
suggestions of mine for immediate activity - please add any others that
you can think of, and we can discuss at our meeting on January 9.

1) Hopefully the DCAT 1.1 work can begin with a review of the
requirements. As a "list" person, I'd probably want to create a list of
potential additions to DCAT and prioritize them in a way that helps the
editors make progress (e.g. easy/obvious, good but harder to define, may
not make it). Please do what *you* think will jump start that work.

2) Before work can get underway on the application profiles deliverable,
we need to define what we mean by profiles and application profiles.
Ruben and I have made a start of a discussion [1] but we need more
voices. You don't have to provide your own definitions if you don't
want, but at least make comments.

Thanks, and we'll post an agenda in the next day or two.

[1] https://www.w3.org/2017/dxwg/wiki/ProfileContext

--
Karen Coyle
kcoyle@kcoyle.net<mailto:kcoyle@kcoyle.net> http://kcoyle.net
m: 1-510-435-8234 (Signal)
skype: kcoylenet/+1-510-984-3600<tel:+1%20510-984-3600>