– DRAFT –
DXWG DCAT subgroup teleconference

https://‌raw.githack.com/‌w3c/‌dxwg/‌andrea-perego-dcat-implementation-report/‌dcat-implementation-report/‌index.html

approva agenda

<riccardoAlbertoni> I am happy with the agenda

AndreaPerego: are you happy with the agenda?

AndreaPerego: a few issues for DCAT2 ratification that need some discussion
… perhaps we can sort them out
… Peter wrote to me adding another agenda item
… pointing to the poll considering the re-charter
… if we are happy with it, we need to submit the re-charter by 31st October
… the final decision should be made tomorrow
… there are a few more editorial changes, but they are not so critical
… how scope is formulated, and so on

SimonCox: I raised the issue about the evergreen standard before
… Philippe said that the evergreen process isn't ready yet
… I'm not totally happy about that
… as undermines the principle of the way we are expecting to work on the DCAT work
… I'd like to have some confirmation of that
… I'm not interested in going through another waterfall style process
… if we are expected to go through that
… I'm not interested in it

<riccardoAlbertoni> +1 not to have the waterfall again for DCAT

<AndreaPerego> alejandra: I also raised the issue about the fact that we have already users' stories.

<SimonCox> For an 'evergreen' process, I would expect to see us work primarily on issues, rather than develop new UCR document as first step

<SimonCox> ... we already have an issue backlog to clear

<AndreaPerego> ... PWinstanley replied saying that the deadlines is only about the deliverables.

<SimonCox> alejandra: for evergreen process, do very frequent working drafts, each targetting a specific issue, which gets us 'evergreen' in practice,

AndreaPerego: what I saw in the new document, the explicit deliverables are DCAT and conneg
… but the UCR is listed among the documents that may be published

https://‌raw.githack.com/‌w3c/‌dxwg/‌e21356361acccddc45472349fd443bbc3954ceb9/‌charter/‌index.html

<AndreaPerego> https://‌raw.githack.com/‌w3c/‌dxwg/‌recharter_2/‌charter/

this was my comment: https://‌github.com/‌w3c/‌dxwg/‌pull/‌1116#issuecomment-538311708

<SimonCox> I would expect to see 'GitHub issue backlog' as one of the inputs to the DCAT3 work

<SimonCox> and remove UCR deliverable for DCAT

indeed, and there shouldn't be a requirement to include the UCR

happy with agenda too

<AndreaPerego> [agenda approved]

approve last meeting minutes https://‌www.w3.org/‌2019/‌09/‌25-dxwgdcat-minutes

<AndreaPerego> +1

<SimonCox> (Still on re-charter - I'm also a bit uncomfortable seeing 'DCAT3' listed - for evergreen it shoudl be DCAT4, DCAT5, DCAT6 ...)

<riccardoAlbertoni> +1

+0 (wasn't there)

<SimonCox> 0 (wasn't there)

Resolved: last meeting minutes approved

implementation report

<AndreaPerego> https://‌raw.githack.com/‌w3c/‌dxwg/‌andrea-perego-dcat-implementation-report/‌dcat-implementation-report/‌index.html

AndreaPerego: summary of the structure
… started with Data on the web best practices
… and removed the non-relevant parts
… used the evidence that we have on the spreadsheet
… and tried to clarify the methodology at the beginning
… what we've done
… we are not listing all that is included in DCAT2
… but only the revisions that need to be verified in terms of implementation evidence
… there is a table with a full list of revisions
… and that section explains that we are not including dcat:Resource and some of the properties as it is an 'abstract' class
… first point is vocabularies using DCAT and DCAT2
… same thing about data catalogues
… for the implementation evidence, for each revision there is the list of implementations as well as possible future implementations
… such as DCAT-AP that is being revised to align to DCAT2
… but also other situations, such as other profiles of DCAT (such as geoDCAT-AP) that have services
… but use a different implementation
… in these cases it is likely that they will be revised to include what is in DCAT2

AndreaPerego: I still need to find out if possible future implementations can be counted
… if they have a high degree of probability
… we have the figures and the summary of the results
… 81 revisions and at least an implementation and a possible future implementation

AndreaPerego: if we look at the charter, we don't need to have all the implementations

<AndreaPerego> From the charter: "In order to advance to Proposed Recommendation, the WG should show that each term in the revised version of DCAT is used in multiple catalogs and related systems. As a minimum, evidence will be adduced that each term has been published and consumed independently at least once, although a higher number is expected for the majority of terms."

ack ;-)

AndreaPerego: 38 revisions out of 81 have at least 2 existing implementations.
… 32 revisions out of the remaining 43 have at least 2 implementations, counting both the existing and the future possible implementations.
… The remaining 11 revisions have 1 future possible implementation, but no existing implementations.

<riccardoAlbertoni> and dct:rights

AndreaPerego: only 11 properties not complying with the requirement

<AndreaPerego> https://‌raw.githack.com/‌w3c/‌dxwg/‌andrea-perego-dcat-implementation-report/‌dcat-implementation-report/‌index.html#analysis

AndreaPerego: properties inherited and the most critical issue is for those properties that are not inherited

riccardoAlbertoni: about the implementation that we have to provide, I'd first check the properties that are more critical
… that we didn't mark as 'at risk'
… the three properties dct:rights

<AndreaPerego> REV32
… dcat:Dataset
… dct:rights

riccardoAlbertoni: dct:PeriodOfTime

<AndreaPerego> REV77
… dct:PeriodOfTime
… time:hasBeginning

<AndreaPerego> REV78
… dct:PeriodOfTime
… time:hasEnd

riccardoAlbertoni: REV77 and REV78

riccardoAlbertoni: if future implementations can be counted or not, it depends on the committment of the future implementation
… are you the editor of any of those future implementations?

AndreaPerego: no
… for services, GeoDCAT-AP there weren't classes
… but we can point to the issues in their repository on github
… they are not going to finish this work before the end of this month

<AndreaPerego> List of issues about the alignment of DCAT-AP with DCAT2: https://‌github.com/‌SEMICeu/‌DCAT-AP/‌issues?q=is%3Aissue+is%3Aopen+label%3A%22dcat+2%22

alejandra: given that some of the profiles are showing their work in progress by moving to DCAT2, that should be quite convincing

alejandra: I was also wondering if you added the implementation reported at https://‌github.com/‌w3c/‌dxwg/‌issues/‌1115
… even if they are work in progress, we might check if they are dealing with the properties that riccardoAlbertoni mentioned

AndreaPerego: all the national extensions in Europe may end up being automatically aligned with DCAT2

AndreaPerego: it'd be good if Philippe joins the call tomorrow
… to ask about these future implementations

the linkedDataPipes are not addressing the 11 properties

AndreaPerego: we are not listing all the (future) implementations that we know of

<SimonCox> maybe change the wording from 'future possible implementation' to 'planned implementation'

this is Simon's endpoint I think: https://‌gnafld.net/‌2016-05/

AndreaPerego: all the European catalogues use DCAT-AP, but we are only listing the European Data Portal

alejandra: will send a PR to your branch with a few editorial changes

<SimonCox> The ones marked 'at risk' can be changed to 'non-normative'

alejandra: what is the process if we don't find implementation of those marked at risk?

AndreaPerego: yes, they need to be removed
… if future possible implementations can count as existing ones, then we are safe

<SimonCox> Or marked 'non-normative' but left in the document.

SimonCox: wording... change to planned implementation
… the things we don't have sufficient implementation evidence that we're marked at risk, I think it is ok to mark them as non-normative and leave them in the document

AndreaPerego: where do we move them?

SimonCox: leave them where they are but putting non-normative

AndreaPerego: when they are inherited, they are listed in the super-class
… only

<SimonCox> and change text from 'is recommended for use' to 'is available for use' for the non-normative properties

<riccardoAlbertoni> "available for use" instead of "recommended" sounds good for me

<SimonCox> My sense is that it is OK for items that were 'at risk' to be changed to 'non-normative' and 'available' (rather than 'recommended') within the spirit of the at-risk warning.

<SimonCox> RRSAgent: generate minutes v2

<AndreaPerego> [meeting adjourned]