DXWG Plenary

19 February 2019

Meeting minutes


<kcoyle> https://‌www.w3.org/‌2017/‌dxwg/‌wiki/‌Meetings:Telecon2019.02.19

<kcoyle> https://‌www.w3.org/‌2019/‌02/‌12-dxwg-minutes

approve last week minutes

proposed: accept minutes

<AndreaPerego> +1

<Makx> 0 not there

<kcoyle> +1


<riccardoAlbertoni> 0 ( i was not there)

<antoine> 0

<roba> +1

Resolved: accept minutes

open action items

kcoyle: there are no plenary open actions, but there are plenty from subgroup activity, some going back some time
… e.g. #135 - #175, #193

roba: #193 is still open, there are conneg issues in discussion and the test suite is in development

kcoyle: discussion on individual open items

<antoine> mail about 244 at http://‌lists.w3.org/‌Archives/‌Public/‌public-dxwg-wg/‌2019Feb/‌0309.html

kcoyle: the issue of profiles has come up in multiple places, but we need to make a decision on action 244, but it is complex and so cannot be done quickly

kcoyle: is it better to break down to smaller pieces?

antoine: it may be that we discuss aspects of it. e.g. #769 discussion would be related to what I'm trying to cover

kcoyle: Peter and I will think about this and come up wiht a plan

antoine: let's keep it on the agenda for a couple of weeks and see if the issues get some focus and then I'll relate these to this work

AndreaPerego: what is the current situation - there is a confusion re: what a resource descriptor is. Some say a distribution, but others say it is between a profile distribution and the profile itself. I think it is like a distribution

<roba> +1 its just a qualified association to an artefact/resource as a dcat:Distribution is

AndreaPerego: we shouldn't complicate the data model
… for me I've not yet seen a strong case for something intermediating between a profile and its distribution. The distribution is the artefact.
… Same with roles - concept or relationship. It can be a relationship

kcoyle: to me this gets to the basic model, and we need to agree this prior to moving forward. We need to bring this down to a vote so we see where people are.

<SimonCox> +1 to looking at concrete examples - I found them very useful in DCAT discussions

<AndreaPerego> +1 to examples

antoine: this is worth a plenary discussion - 30+ mins. A lot of discussion happens without looking at concrete examples, and from those I can illustrate why I think there is a need for the intermediate level. I think that AndreaPerego and I agree, but perhaps others don't

kcoyle: who can create sensible readable examples?

roba: there are examples, Makx showed one at the F2F.

<SimonCox> Put example snippets into the issues - don't rely on people going out to the repo

<SimonCox> at least put specific links into the issue thread

<roba> https://‌w3c.github.io/‌dxwg/‌profilesont/‌examples/

roba: Examples are through the profiles ont doc

kcoyle: we need links. issue #529. Put snippets into the issue
… let's do the discussion there

<roba> https://‌www.w3.org/‌TR/‌dx-prof/

antoine: I'm volunteering to add these examples in my docs (examples -> URIs)

kcoyle: are you pointing to roba examples?

antoine: I think I have, I'm trying to link the prof vocab to other things

antoine: I wouldn't put everything in #529 - it is large and will become unwieldy
… but we probably shouldn't have another issue for everything. I can refer to the section of my alignment from #529

<Zakim> AndreaPerego, you wanted to briefly comment on PROF examples - e.g., https://‌www.w3.org/‌TR/‌dx-prof/#ex-1-dcat-ap-described-using-profile-ontology-in-rdf-turtle

kcoyle: do we deal with this as a whole, or fragments?

AndreaPerego: some examples in the prof onto are not clear. the one I linked to is about DCAT, but it seems misaligned with the discussion

<antoine> +1 to Andrea's point: this is the sort of inconsistency that would make work on it hard

<roba> yes - that example is broken - the one in the github is right :-(

kcoyle: we may need new examples

roba: AndreaPerego picked up an incorrect example. I will look at it - it is an editorial problem
… #529 was originally an alignment to DCAT, but the prof voc doc needs to emphasise the relation

<SimonCox> This whole discussion was triggered when Makx showed how they had shoe-horned profile descriptions into a DCAT structure. roba picked this up and has tried to clarify it or explore it. We do need to be clear about whether we want Profile descriptions dependent on DCAT in this way ...

antoine: this is not about one specific example, there is no example of prof:hasArtifact in the prof onto doc
… perhaps it is good for roba to sort this prior to other work on this

<SimonCox> message to roba - it doesn't work to refer only to examples in the repo, when in fact they are inconsistent

Action: roba to clean up examples

<trackbot> Created ACTION-295 - to clean up examples [on Rob Atkinson - due 2019-02-26].

<SimonCox> ... and then put snippets into the issues (they will be dated by the issue comment)

kcoyle: we have a way forward - look at examples, then work out how to move forward with the diagrams antoine is making

antoine: in the meantime I'll focus on specific issues, and when roba updates the examples i need to confirm my understanding of the ontology
… being involved in the paper on prov I'm familiar with this area, but need to double check after the example update


<alejandra> and there is a PR to implement the change from ontology to vocabulary: https://‌github.com/‌w3c/‌dxwg/‌pull/‌768

kcoyle: Next Item: the vote in favour of renaming the profiles ontology to a profile vocabulary

<roba> https://‌joinup.ec.europa.eu/‌rdf_entity/‌http_e_f_fdata_ceuropa_ceu_fw21_f6f27f059_bf785_b4d7d_bb602_b6448aab73bd5 is a dcat:Distribution - and is in fact a Resoruce descriptor .. it is missing the artefact link: which shoul;d be https://‌joinup.ec.europa.eu/‌sites/‌default/‌files/‌distribution/‌2015-06/‌dcat-ap_revision_draft_5_v0.05.docx

Jaroslav_Pullmann: a vocabulary is normally a set of qualified values, whereas the ontology is a model
… what is the rationale?

<roba> +1 to jaro

<Jaroslav_Pullmann> https://‌www.w3.org/‌TR/‌odrl-vocab/

<Jaroslav_Pullmann> https://‌www.w3.org/‌TR/‌odrl-model/

<roba> prof defiens a namesapce and doesnt "combine" - i.e. profile other vocabularies

kcoyle: from a comment received - in some peoples' minds (W3C is equivocal) an ontology is in a single namespace, but a vocabulary combines them

Jaroslav_Pullmann: there is no hard rule, but this was my understanding

<riccardoAlbertoni> +1 to AndreaPerego

AndreaPerego: to add confusion, what Jaroslav_Pullmann calls a vocabulary is often called a controlled vocabulary, but in the semweb world there is a tendency for situations that are heavily axiomatised the term 'ontology' is used, but for lighter semantics the term 'vocabulary' is used

<kcoyle> https://‌lov.linkeddata.es/ linked open vocabularies

AndreaPerego: but this distinction is very domain dependent

<alejandra> +1 to AndreaPerego's description, which also differs from the hierarchy of controlled vocabulary/thesaurus/ontology

AndreaPerego: my vote in favour of using 'vocabulary' is because of the low level of axiomatisation

roba: I'm agnostic.
… different sides of the same coin
… to be consistent with DCAT we might as well use the term 'vocabulary'

Resolved: rename Profiles Ontology to Profiles Vocabulary

<kcoyle> https://‌github.com/‌w3c/‌dxwg/‌issues/‌747#issuecomment-464371792

kcoyle: another issue for voting: should we include prof in the next WD. 12 votes in favour

roba: were a set of roles defined in another doc, moving those into the ontology doc is fine.

kcoyle: we want them listed in the doc so that people actually see them - it is not about namespace

roba: we cannot action that until we know what they look like. we might need expert advice on OWL DL to ensure consistency

antoine: I think they can be in the same namespace for the moment

<AndreaPerego> +1 to use the same namespace

<roba> +1

antoine: I think we should keep things simple

<riccardoAlbertoni> +1 to antoine's suggestion

<Zakim> AndreaPerego, you wanted to ask a clarification why properties will prevent the use of multiple roles

antoine: this depends on the conclusion to the discussion on how to present roles

AndreaPerego: what is the reason? why can we not have multiple relationships?

roba: I suppose you can repeat the relationship
… unusual pattern of A with 5 relationships to B, but it can be done

Resolved: Include a core set of roles in the next working draft of the Profiles Vocabulary

<SimonCox> I see no problem with multiple roles on one relationship

kcoyle: I will include a link to this resolution in the issue, so that everything is connected.

<SimonCox> the association class is an artefact to handle exactly this kind of thing

kcoyle: Next Item: there are many open issues - many seem to be dormant

<roba> but ResourceDescriptor is the association class - thats the pattern we currently have...

kcoyle: or they are done but not closed

<kcoyle> https://‌github.com/‌w3c/‌dxwg/‌issues/‌3

kcoyle: if possible

<roba> (i looked at each.. +1 to close them all)

kcoyle: Let's work through these quickly

antoine: it has been merged.

<SimonCox> try this view ... https://‌github.com/‌w3c/‌dxwg/‌issues?q=is%3Aissue+is%3Aopen+sort%3Aupdated-asc

<kcoyle> https://‌github.com/‌w3c/‌dxwg/‌issues/‌7

alejandra: I think we can - but there is a list that Makx put together , let's check that they are on the spreadsheet for the implementation analysis

kcoyle: can Makx copy to the wiki page?

<kcoyle> proposed: close #7 after Makx copies his info to the wiki page

<alejandra> +1

<riccardoAlbertoni> +1


<SimonCox> +1

Resolved: close #7 after Makx copies his info to the wiki page

<kcoyle> https://‌github.com/‌w3c/‌dxwg/‌issues/‌9

Action: Jaroslav_Pullmann will close issue #9

<trackbot> Created ACTION-296 - Will close issue #9 [on Jaroslav Pullmann - due 2019-02-26].

<kcoyle> https://‌github.com/‌w3c/‌dxwg/‌issues/‌14

Resolved: closed #14

antoine: if we close them because they are done, then add a note so that we know what was done

<kcoyle> https://‌github.com/‌w3c/‌dxwg/‌issues/‌32

<kcoyle> https://‌github.com/‌w3c/‌dxwg/‌issues/‌37

alejandra: I'll check issue #37

Action: alejandra check issue #37

<trackbot> Created ACTION-297 - Check issue #37 [on Alejandra Gonzalez Beltran - due 2019-02-26].

<kcoyle> https://‌github.com/‌w3c/‌dxwg/‌issues/‌38

<antoine> https://‌w3c.github.io/‌dxwg/‌ucr/#ID50

antoine: I think I've found something like this -

Action: Jaroslav_Pullmann issue #38

<trackbot> Created ACTION-298 - Issue #38 [on Jaroslav Pullmann - due 2019-02-26].

please use https://‌github.com/‌w3c/‌dxwg/‌issues?q=is%3Aissue+is%3Aopen+sort%3Aupdated-asc to find issues to close


<alejandra> thanks! bye

<AndreaPerego> Thanks, bye bye!

<Jaroslav_Pullmann> thanks & bye!

Summary of action items

  1. roba to clean up examples
  2. Jaroslav_Pullmann will close issue #9
  3. alejandra check issue #37
  4. Jaroslav_Pullmann issue #38

Summary of resolutions

  1. accept minutes
  2. rename Profiles Ontology to Profiles Vocabulary
  3. Include a core set of roles in the next working draft of the Profiles Vocabulary
  4. close #7 after Makx copies his info to the wiki page
  5. closed #14
Minutes manually created (not a transcript), formatted by Bert Bos's scribe.perl version 2.49 (2018/09/19 15:29:32), a reimplementation of David Booth's scribe.perl. See CVS log.


Succeeded: s/topic: open action items//

Succeeded: s/prof as artefact/prof:hasArtifact

Succeeded: s/inconsistens/inconsistent/

Succeeded: s/Max/Makx