DCAT 2018-07-26

26 July 2018

Meeting Minutes

<SimonCox> Minutes of last meeting https://‌www.w3.org/‌2018/‌07/‌19-dxwgdcat-minutes

proposed: accept Minutes of last meeting https://‌www.w3.org/‌2018/‌07/‌19-dxwgdcat-minutes

<PWinstanley> +1


<DaveBrowning> +1


<Jaroslav_Pullmann> +1

<SimonCox> +1

Resolved: Minutes of last meeting https://‌www.w3.org/‌2018/‌07/‌19-dxwgdcat-minutes

<SimonCox> Catalogues in which dataset is an undifferentiated set of files https://‌github.com/‌w3c/‌dxwg/‌issues/‌256 Consolidated PR https://‌github.com/‌w3c/‌dxwg/‌pull/‌295 or see draft https://‌rawgit.com/‌w3c/‌dxwg/‌dcat-dataset-relations-simon/‌dcat/‌index.html#Property:resource_relation ready to merge

DavidBrowning will merge

Minor issue https://‌github.com/‌w3c/‌dxwg/‌issues/‌64

<SimonCox> https://‌rawgit.com/‌w3c/‌dxwg/‌dcat-issue64-simon/‌dcat/‌index.html#classifying-dataset-types

SimonCox: discuss complete docu to close https://‌github.com/‌w3c/‌dxwg/‌issues/‌64
… Question 1: legal for RDF to have multiple types - but should example separate?

<PWinstanley> I think we should have a simple case and a more complex case as illustrations

SimonCox: Question 2: how to handle when type is not a resolvable URI - couple of examples shown - what do people want

<PWinstanley> I'm slightly allergic to blank nodes too

DaveBrowning: quite like example
… blank nodes (squirming noises...) can live with

SimonCox: in as PR \

<Jaroslav_Pullmann> ah, you are possibly talking about those constructs: 'dct:type[ rdfs:label "Text" ; dct:source "DataCite resource types" ; ] ;' -> what is the minimal model/predicates to go into the node to become a "valid" content type indicator?

DaveBrowning: if use blank nodes should be in a coherent way

<PWinstanley> I think we need to be pedagogical too - people need to be able to learn from these documents. We are not writing them (only) for experts

Action: DaveBrowning to smooth out examples, including dct:type example, as part of editorial sweep

<trackbot> Sorry, but no Tracker is associated with this channel.

roba: is this a qualified forms example - and how do we "be coherent" in approach whenever it comes up?
… architectural choice - can you rely on object URI dereferencing to get metadata or do we inject nodes to hold metadata?

SimonCox: there is no answer at this point..

<PWinstanley> i think we can tidy with editing

SimonCox: can we merge and tidy up in editing

<alejandra> #Hi all, I'm travelling and unable to connect to audio

<SimonCox> Proposed: merge https://‌github.com/‌w3c/‌dxwg/‌pull/‌308

<SimonCox> +1

+0 (not over details)

<DaveBrowning> +1

<SimonCox> ... and tidy up doco in editting later per action on DaveBrowning

<PWinstanley> +1

Resolved: merge https://‌github.com/‌w3c/‌dxwg/‌pull/‌308

Action: DaveBrowning to merge https://‌github.com/‌w3c/‌dxwg/‌pull/‌308

<trackbot> Sorry, but no Tracker is associated with this channel.

Initial examples of use of PROV for dataset Project Context

<SimonCox> https://‌github.com/‌w3c/‌dxwg/‌issues/‌71

<SimonCox> https://‌github.com/‌w3c/‌dxwg/‌issues/‌71#issuecomment-404059767

<SimonCox> https://‌github.com/‌w3c/‌dxwg/‌issues/‌71#issuecomment-404059767

<SimonCox> Use of prov:wasGeneratedBy ?

<SimonCox> Alternative options:

<PWinstanley> With Simon's proposed Project Ontology, that is only one of several in this area: http://‌citeseerx.ist.psu.edu/‌viewdoc/‌download?doi= So we need to be careful not to hard wire that into the DCAt work

<SimonCox> 1. recommend use of prov::wasGeneratedBy, with potential side-effects

<SimonCox> 2. 'clone' wasGeneratedBy in dcat: NS, and manage alignemtns as much as desired

<AndreaPerego> +1 to using prov:wasGeneratedBy

<PWinstanley> +1 to option 1

roba: can we just warn people of prov: and its potential entailment baggage?

<AndreaPerego> +1 to guidance on using prov:

<alejandra> I'm not clear about the potential side-effects, but in principle I favour to reuse prov:wasGeneratedBy rather than 'clone' it

PWinstanley: reason for using prov: ... if we clone our own we end up with nuanced versions and this "spins away" from Linked Data towards separet dialects of RDF

<SimonCox> Proposed: recommend use of prov:wasGeneratedBy to give dataset context

DaveBrowning: agrees.. prov is a good tool or not - if need something very different use it

<AndreaPerego> +1 to PWinstanley & DaveBrowning

<SimonCox> +1

<DaveBrowning> +1

<Jaroslav_Pullmann> I do not see troubles with the axiomatization of prov:wasGeneratedBy, since prov:Entity is abstract; but is there a way to express the "funding" Action?


<PWinstanley> +1

<Jaroslav_Pullmann> +1

Resolved: recommend use of prov:wasGeneratedBy to give dataset context

<SimonCox> I would say that funding is a property of the activity (one step removed from the dataset)


<SimonCox> prov is very activity-oriented !

<Jaroslav_Pullmann> ok and the founder is an Agent

roba: and this is another qualified forms problem...

<SimonCox> ... please look at Project ontology https://‌dr-shorthair.github.io/‌ont/‌project/

roba: a straw man - no intention this is a recomendation - possible Note

<SimonCox> in this proposal project is a specialization of prov:Activity

<Jaroslav_Pullmann> @Simon, the Project ontology is really interesting - I could imagine there are similar initiatives to describe (research) projects, e.g. http://‌mayor2.dia.fi.upm.es/‌oeg-upm/‌index.php/‌en/‌ontologies/‌81-research-proj-ontologies/‌index.html

<SimonCox> yes Jaroslav_Pullmann - also see list here https://‌dr-shorthair.github.io/‌ont/‌project/#informative-references

<PWinstanley> yup!!

<Jaroslav_Pullmann> @Simon: "Project is an Activity", fits fine with the definition "something that occurs over a period of time and acts upon or with entities"

roba: how much should be try to provide options to plug identified holes in available vocabularies to meet identified requirements:

<Jaroslav_Pullmann> .. just out of interest: the the project funder (e.g. EC) would be related via the project ontology, e.g. proj:hasSponsor ?

DaveBrowning: talks about related work (we publish datasets continually so project is maybe not the best fit but something playing the same role....) .. quite like it - boss would expect model closer to specific needs

SimonCox: I work for an org that works via projects - others work on continuous production of data, governments may work on services
… answering Jaro - previous projects have been through lens of academic grants (holding sides and chortling...)
… shows pattern works not just for data generation

<SimonCox> See Figure 3. for funding pattern

<SimonCox> (answer to Jaroslav_Pullmann )

SimonCox: back to issue - proposed "wasGeneratedBy" - range is provActivity - time bounded things

<Jaroslav_Pullmann> @Simon, thanks, the "funding" is an important aspect of the provenance (~ influence)

roba: potentially time-bounded, but not forced to be so?

SimonCox: examples in PROV literature are all "instantaneous"

DaveBrowning: looked into this too... more later this year

<AndreaPerego> https://‌ec-jrc.github.io/‌datacite-to-dcat-ap/#alignment-issues-agent-roles

AndreaPerego: evidence to support the apprach

<PWinstanley> bye!!

<Jaroslav_Pullmann> bye, thanks for an interesting meeting!

Summary of Action Items

  1. DaveBrowning to smooth out examples, including dct:type example, as part of editorial sweep
  2. DaveBrowning to merge https://‌github.com/‌w3c/‌dxwg/‌pull/‌308

Summary of Resolutions

  1. Minutes of last meeting https://‌www.w3.org/‌2018/‌07/‌19-dxwgdcat-minutes
  2. merge https://‌github.com/‌w3c/‌dxwg/‌pull/‌308
  3. recommend use of prov:wasGeneratedBy to give dataset context
Minutes formatted by Bert Bos's scribe.perl version 2.41 (2018/03/23 13:13:49), a reimplementation of David Booth's scribe.perl. See CVS log.


Succeeded: s/please enter details/we publish datasets continually so project is maybe not the best fit but something playing the same role/

Succeeded: s/examples/examples in PROV literature/