– DRAFT –
Dataset Exchange Working Group Teleconference

kcoyle: approval of last week's minutes

<SimonCox> As I was not present, no comments

<SimonCox> https://‌www.w3.org/‌2017/‌10/‌23-dxwg-minutes

<kcoyle> https://‌www.w3.org/‌2017/‌10/‌23-dxwg-minutes

<annette_g> ditto

Resolved: approve last week's minutes

kcoyle: Open Actions:
… Makx sent in on action #18

<kcoyle> Can we consider Action 18 completed?

<Makx> yes as far as i am concerned

<Makx> it's just that I did not know where to put the stuff

kcoyle: Action #18 will be closed. Makx sent in a use case that is now in the wiki doc but needs to be moved to the definitive doc

<Makx> I put it in https://‌www.w3.org/‌2017/‌dxwg/‌wiki/‌Use_Case_Working_Space

<SimonCox> There are too many routes to add UC's!

kcoyle: Jaroslav_Pullmann will do this

<Makx> At https://‌www.w3.org/‌2017/‌dxwg/‌wiki/‌Use_Case_Working_Space#Quality_information

SimonCox: this highlights a problem - the multiple routes to creating a UC

<Makx> It's also that I don't know how to fork GitHub

kcoyle: we had decided using issues in github

Action: Jaroslav_Pullmann move Makx's new use case into UCR draft

<trackbot> Created ACTION-52 - Move makx's new use case into ucr draft [on Jaroslav Pullmann - due 2017-11-07].

kcoyle: we can also mark documents on the home page as being obsolete and this should make it more obvious that it is not the route to use

<Makx> OK, my fault then

kcoyle: Any other completed actions?

Jaroslav_Pullmann: I worked on content for Action #48

<Jaroslav_Pullmann> https://‌www.w3.org/‌2017/‌dxwg/‌wiki/‌General_versioning_considerations

<roba> wiki states this "!!This document was for draft discussions and is now inactive. Further development has now moved to structured documentation on the github at https://‌github.com/‌w3c/‌dxwg/‌blob/‌gh-pages/‌ucr/‌index.html"

Jaroslav_Pullmann: the wiki page I created contains the graphics

<Makx> Yes roba I saw that after I addded the UC. My fault.

<roba> maybe it ought to say "submit an issue" - what about people outside the group - email ?

Jaroslav_Pullmann: this is a first, quick analysis of the levels we should consider, but they have different lifecycles and there needs to be some profiling

kcoyle: I will put this in with the 'versioning' section in the F2F
… I will also get together with Caroline and annette_g (who will help me chair) and we need an agenda that fits with the times that people accessing remotely will be present

<Caroline> s/

kcoyle: at the F2F we should complete all the requirements and allocate them to deliverables. Can people look at the categories roba developed, and Alejandra's doc

Jaroslav_Pullmann: I will contribute further to the topic - to determine the levels that we have and to relate these to the profiles. This refers to the wiki page

kcoyle: I hope that we will have our meeting next week on Tues (the F2F is Thu and Fri) as it gives us a last minute chance for planning

kcoyle: Moving to requirements:
… #614 'provide fine-grained metadata for DCAT distributions'

roba: I grouped a range of disparate special cases requirements all about fine grained metadata for distributions, and I interpreted specific requests about essentially being the same issue, viz how can communities provide models/ Are there any common attributes that should go into DCAT, and what is to go in profiles

<roba> +1

<roba> i think deal with the general case

kcoyle: roba do you think we should consider as a group, or just go ahead?

<roba> if there is something not easy to catch then spin out a special case req

roba: I think I would gather similar issues, and if there is anything unique then that goes in a specif requirement

roba: we should dive into solution space and see if there is a canonical connection point for detailed metadata, or do we leave it for the application profile
… We have made progress by introducing the AP approach to allow communities flexibility. But, we need a clear mechanism for extension that uses a pattern consistent with the ways we are implementing elsewhere in the DCAT
… We need a generlisable solution for incorporating 3rd party vocabs

Jaroslav_Pullmann: is the wording clear enough for general users
… Should we state more clearly that this is a metareuirement of the ability to define speccific extensions

kcoyle: This is one of those areas where a short ....

kcoyle: the heading and the more detailed description are different.

<annette_g> +1 to KCoyle

kcoyle: the heading should read "...provide for fine grained metadata .."

<roba> i have no problems with that :-)

kcoyle: so essentially we can make that change
… Any other comments?

<roba> update the spreadsheet, and i will transcribe to UCR in a batch later (after groupings sorted out)

<roba> examples are described partially by links back to UC

Jaroslav_Pullmann: shouldn't we provide examples of the dynamic distributions or the domains we are about to attach via profiles? What types of fine-grained metadata will come in through these extensions?

<SimonCox> FWIW link to editor's draft https://‌w3c.github.io/‌dxwg/‌ucr/#RID12

Makx: I am with Jaroslav_Pullmann . If I look at #614 and #615 they don't say much. There are cases where if the work we are doing in the EC came across (dynamic datasets, service bsed profiles) . We need to be more specific detiling the things people are encountering

annette_g: We need to thingk about the actual UC. The desfription is for a general feature. If we knew the need we can respond and make the solution more specific

roba: as SimonCox pointed out, on the editors draft there are examples. there are links to the UCs . There are examples in Void, and other linked data vocabs there are examples. fine-grained is a catch-all that we are not considering as a first-class DCAT component
… there is always a need for fine-grained detail, but this doesn't stop us from standardising widely used types of fine-grained metadata

<annette_g> +1 to defining the requirement as enabling usage of classes/terms from other vocabs.

<Makx> +1 to roba

roba: There needs to be the general mechanism for handling things that are not in core DCAT

<SimonCox> +1 to clear extension points + patterns

Jaroslav_Pullmann: the focus is to provide defined extension points and guidance on extending where the vocabu in the next evolution will be described as part of DCAT. The proposal of extension points for profile work is the focus of this requirement

kcoyle: can someone provide new wording?

<kcoyle> Define a way to attach finer grained metadata for dcat:Distribution instances associated to a dcat:Dataset

<roba> let me try

<Jaroslav_Pullmann> +1

kcoyle: can this either be rewritten on the fly, or else someone take an actions

<roba> Define a way to attach finer grained metadata where DCAT does not specify an appropriate property.

<annette_g> Define a way to attach metadata from external vocabularies to dcat:distributions or dcat:datasets.

kcoyle: roba and annette_g have provided alternative wordings

<roba> happy with Annettes

<SimonCox> Sorry - I have to drop off now to catch a bus

<roba> i too have to dissappear in the next minute or too :-(

kcoyle: we have a new wording proposal from annette_g for #614

<roba> just noting this covers 6:13 as well!

<kcoyle> PROPOSED: accept requirement 6.14 using Annette's wording

<Jaroslav_Pullmann> this seems to be quite generic, see UC ID26

Jaroslav_Pullmann: this is quite generic. The UC on providing extension points and examples was there at the beginning. We need to be clear about what is to be in new DCAT and what will be in the AP. We need to be cautious about being too abstract/generic.

<roba> sorry i have to go.

annette_g: I see nothing wrong with being more specific, but it requres those who want it to define what they are looking for

kcoyle: how do you feel about the UCs?

annette_g: we need to make DCAT extendable, but I don't know why we need to be more specific

Makx: We can get the specific examples from the UC, but (agreeing wiht PW) there is a need to give people examples. So, going back to the UC (e.g. the one on web services and interfaces) we can be specific about what is needed
… maybe the requirement should be made more concrete

kcoyle: I am looking to see if the web services became .... for some UC there are specific requirements
… It seems we need an action to create a specific requirment for those where one does not currently exist

Jaroslav_Pullmann: I think this is a good example of what Makx mentioned - e.g. web servisces might be http , but they might be other protocols. We need to be able to describe in a DCAT-compatible manner the provision of a web service by another protocol
… maybe '614 should be decided at the end, being a waste bin for what we don't care about right now, and then we can provide examples

kcoyle: I propose we put aside until we can determine if we really need it

kcoyle: #620
… this seems more specific

<kcoyle> https://‌w3c.github.io/‌dxwg/‌ucr/#RID18

Jaroslav_Pullmann: can data quality vocbulary be used here?

<annette_g> antoine and riccardo are primary authors of DQV

Makx: we have done some work on textual quality information. but we haven't looked into the rest of DQV which is about metrics etc.

<annette_g> DQV doc intro says "The Data Quality Vocabulary (DQV) presented in this document is foreseen as an extension to the DCAT vocabulary"

kcoyle: there are a number of UC from Alejandra about modelling quality and performance
… any comments?

Jaroslav_Pullmann: we were talking about data quality modelling. we need to assess what is mandatory and is a candidate for part of the new DCAT, and what needs to be left over for optional work that might be mandatory for specific communities but irrelevant for others
… we need to work out the quiality aspects that are general

annette_g: The way DQV was conceived was to make it flexible
… it is more about giving people what they need for their domain.
… This should address Jaroslav_Pullmann concerns

Makx: I agree with annette_g , but DQV is young and not widely used. It is also a W3C note and so doesn't need use. We need to talk to Antoine to see if it is used in Europeana

annette_g: Maybe for a requirement we don't need to focus on DQV, but we need to consider quality

kcoyle2: Can we consider the proposal for considering quality as part of DCAT 1.1?

<annette_g> +1 for voting, +1 for the requirement

<kcoyle2> PROPOSED: accept requirement 6.20

<annette_g> +1

+1

<DaveBrowning> +1

<Makx> +1

<Caroline> +1

<Jaroslav_Pullmann> +1

Resolved: accept requirement 6.20

Resolved: defer requirement 6.14

<Makx> OK bye bye

<Jaroslav_Pullmann> bye!

kcoyle2: a first draft of the F2F will be available shortly

<AndreaPerego> s/s\///