– DRAFT –
Dataset Exchange Working Group Teleconference

<PWinstanley> chari: PWinstanley

<PWinstanley> proposed: accept minutes https://www.w3.org/2020/01/26-dxwg-minutes

<PWinstanley> https://www.w3.org/2021/01/26-dxwg-minutes

Admin

<AndreaPerego> +1

<PWinstanley> +1

<plh> +1

<riccardoAlbertoni> +1

PWinstanley: minutes of previous meeting (Jan 26); meeting after that was cancelled

+1

<Rachel> +1

<Ana> +1

<antoine> 0 (absent)

Resolution: accept minutes https://www.w3.org/2021/01/26-dxwg-minutes

Resolution: approve minutes of Jan 26

subgroup meetings

PWinstanley: mainly dcat

<annette_g> +1 fwiw

<riccardoAlbertoni> https://github.com/w3c/dxwg/pull/1292

riccardoAlbertoni: DCAT update: advancing the work on versioning; adding "series"; got feedback on whether class series should be subclass of resource
… deciding to make it a subclass of dcat:Dataset

<riccardoAlbertoni> https://github.com/w3c/dxwg/pull/1295

riccardoAlbertoni: versioning, AndreaPerego has a proposal; it is an evolution of first version of DCAT

PWinstanley: a response to feedback

riccardoAlbertoni: yes, after feedback we added new terms for versioning

PWinstanley: got response from draft that there is a desire for stronger guidance; we've been using a light touch because such a wide field of useres
… we might review this in the future, to give a more focused guidance

riccardoAlbertoni: Important to do these things incrementally, starting with a simple solution then adding
… not taking all the terms from PAV

<kcoyle> s/P?/PAV

PWinstanley: PAV is no longer being maintained, so need to go beyond that
… are the other things that we can take an idea from but provide foundation to take it forward

<antoine> ok question is: will you include PAV terms (some) and then create 'official versions of them' in the DCAT namespace?

<antoine> a bit like we've done in DQC for DAQ

<antoine> ok thanks this sounds good!

riccardoAlbertoni: we are minting new terms that are equivalent to PAV in the DCAT namespace

<annette_g> This all makes me happy.

annette_g: glad that we're putting versioning terms into DCAT

PWinstanley: do you have use cases?

annette_g: not sure

PWinstanley: do we have enough horsepower in the group? Fewer active members

Rachel: Intend to join the subgroup, catch up with what's going on, need more review of the work

PWinstanley: Don't need to hold back; feel free to join and listen

Rachel: Not yet implementing DCAT in work

annette_g: Question about current version term - is this to say this is a more current version and this points to it

PWinstanley: with PAV you have two options: one points to previous, one says this is current

AndreaPerego: what we use is the version chain and the version hierarchy; chain is links to previous/next versions
… the hierarchy from abstract resources to versions
… for DCAT3 there is a URI, then we have working drafts, etc., so the relationship between DCAT3 (unversioned) and the versions
… current version points to the latest stable version (which we don't have yet)
… the section on versioning has graphs that explain how these work

annette_g: is there a use case for current version that would not run into a problem of maintaining versions over time

AndreaPerego: yes, when you have versions of a dataset and they are revised over time you get links to a previous version
… but if you have a URI that points to the latest you have to update the link to point to the new version when those happen
… in DCAT you can decide whether to use one or both of these types of versions

annette_g: it would be good to point out that uri for latest is always correct

antoine: which channel to use for detailed comments?

PWinstanley: probably github

riccardoAlbertoni: yes, github. pull request not yet merged so comments can be added to PR

antoine: PR is pretty complex at this point, and it's hard to find specific issues

riccardoAlbertoni: PR will be merged soon; so probably best to comment in issues

AndreaPerego: prefer github issues
… we have to open issues about the versioning section; the first is related to specifying backward compatibility to previous versions

<AndreaPerego> Backward compatibility: https://github.com/w3c/dxwg/issues/1258

<AndreaPerego> Version delta: https://github.com/w3c/dxwg/issues/1271

AndreaPerego: next - how to specify version delta; should we provide a machine-actionable version delta? needs more discussion

antoine: should create new issues for specifics? (yes)

github housekeeping

PWinstanley: we have a lot of open issues for all projects, and it's difficult to understand scope of work
… some need to be closed; some need to be deferred. Need to deal with those that relate to more than one work product

AndreaPerego: That's right. We have a lot of issues that have had no discussion for years. We have over 200 open issues. Need to decide what to do with them

PWinstanley: close those that are no longer relevant; for those with more than one deliverable decide if there is a primary

antoine: I am guilty for stale issues for profile guidance. This is one reason why I was against splitting the repo's; but the profile vocabulary is working well
… but we could move everything relevant into profile guidance

PWinstanley: over next few weeks look things over individually

<antoine> kcoyle: some time ago I did a list of open issues

<PWinstanley> kcoyle: some time ago, I did a list of open issues

<antoine> ... I will find that again

plh: not sure how to unarchive an archived repo; can use labels on the issues. archiving is a last resort when no longer working on it

PWinstanley: so, characterize, don't archive
… put this on the agenda for next plenary

<Rachel> thank you

<Ana> thank you, bye!

<AndreaPerego> Thanks, bye bye