Meeting minutes
admin
<PWinstanley> proposed: accept minutes https://www.w3.org/2020/02/11-dxwg-minutes
<roba> +0
<riccardoAlbertoni> +0 ( I was not there)
+1
<PWinstanley> +1
<alejandra> +1
<Makx> 0
<ncar> 0
Resolution: accept minutes https://www.w3.org/2020/02/11-dxwg-minutes
DCAT version 3
PWinstanley: splitting of github repo has resulted in some polarized views, some proposals for a staged approach. Antoine was pointing out that it's going to take quite a bit of effort if we do split.
roba: Antoine quoted a message from the email archives but not my response, which doesn't appear in the email archive. I wondered if anyone else had seen issues with messages not appearing in the archives.
PWinstanley: That's in issue 1216
<roba> https://github.com/w3c/dxwg/issues/1216
ncar: I have no problem with splitting the repository. I understand why we had them all together. If there's no or little overlap, then I don't see why we shouldn't split those. (reiterates Karen's analysis)
PWinstanley: simon, you have a similar view, right? and rob?
SimonCox: yes, keeping conversations distinct and keeping builds distinct is best practice.
… also, release tags and such make more sense
riccardoAlbertoni: my only concern is references to github issues that already exist. My recommendation is to leave DCAT in the repo and split out the others.
PWinstanley: that would be leaving all the content in the current github and to set up clones for prof and conneg work?
riccardoAlbertoni: I think it's important to keep DCAT in place, would like to hear from others.
SimonCox: If some of the other docs have links in the tracker, the links stay in there, they don't break because you've closed it.
<SimonCox> When you close an issue, you can still link to it. So the trick would be to put a forward pointer at the bottom of the issue conversation.
ncar: I'm not sure we can leave DCAT in place rather than close it out.
<ncar> An archived repo: https://github.com/CSIRO-enviro-informatics/Place-Names and it's successor: https://github.com/GeoscienceAustralia/Place-Names/
riccardoAlbertoni: what I'd like to avoid is that we have many repos about DCAT.
<ncar> SHACL has many repos in the one space: https://w3c.github.io/data-shapes/
PWinstanley: the question is whether we stick with what we've got or split it. The details of how we split need to be addressed downstream. We just have to be sure we leave the trail in an appropriate state for others to follow. I think we have enough people to make a decision.
… we had a poll, but the discussion kept rolling along.
SimonCox: is it feasible to leave it up to each subgroup?
<alejandra> Nic - those are not several repos for data shapes - it is one repo with many folders
PWinstanley: I think there needs to be a certain amount of coordination.
<alejandra> https://github.com/w3c/data-shapes/
<SimonCox> I think the proposal was to leave DCAT on the DXWG repo, right?
oops, I've been scribing roba as simoncox.
PWinstanley: andrea, any point you want to make?
AndreaPerego: I don't have enough info about what's been discussed.
AndreaPerego: the concern I have is the issue with the issues. Also thinking what will happen to the linked issues.
… If we point to the previous repository from the new one.
<PWinstanley> ack \
roba: I think it's up to the DCAT group how they want to handle it. As long as there's an opportunity to transfer anything that's a live issue, that should work well.
PWinstanley: annette_g
<PWinstanley> annette_g: I'm neutral about this - I can see merits in both approaches
alejandra: I'm in favor of splitting because it will make things easier in the end. for bulk closing issues, I think we first need to move them to the new repos. It would be good to prefix the repo names with the working group name. Another group that splits their repos is json-ld.
<alejandra> https://github.com/w3c?utf8=%E2%9C%93&q=json-ld&type=&language=
alejandra: We need to test how we can move things.
<Makx> i´m neutral
PWinstanley: Ana, Makx, anything?
<alejandra> especially check how to move issues
<Ana> i'm neutral
<alejandra> I wasn't aware of that other repo - can you send the link Nic?
ncar: we have another repo for the group, conformance reports for conneg. I look for guidance for where to put that one. We do already have two repos, so more wouldn't be a big deal.
PWinstanley: let's try a simple proposal
<alejandra> I think it would be good to prefix them with 'dxwg' (I think that wasn't in the minutes from my previous comment)
<ncar> Here is the ConnegP Test Suite repo: https://github.com/w3c/prof-conneg-testing
Proposed: that we split up the repos so that we have one repo per deliverable, DCAT and Conneg.
<ncar> +1
<PWinstanley> +1
+0
<riccardoAlbertoni> +1 (if we keep dcat where it is... )
alejandra: I'm in favor of voting for this, but shouldn't we clarify the details?
PWinstanley: I suggested that we work out the details after we agree to split generally.
… do you agree?
alejandra: yes, but I would put as Riccardo did. I think it would be better to list out the details.
PWinstanley: please type it up
<alejandra> Proposed: that we split up the repos so that we have one repo per deliverable, DCAT and Conneg, considering that DCAT stays in the same repo and all repos are prefixed with dxwg
<alejandra> +1
<riccardoAlbertoni> +1
<Ana> +1
<roba> +1
<PWinstanley> +1
+0.2
<Makx> +1
<ncar> +1
<SimonCox> +1
roba: provided it doesn't prevent us spinning of prof into its own repo.
<alejandra> I agree that PROF would be in another repo too
<AndreaPerego> 0
PWinstanley: right, it's not stopping anything.
Resolution: that we split up the repos so that we have one repo per deliverable, DCAT and Conneg, considering that DCAT stays in the same repo and all repos are prefixed with dxwg
<alejandra> here the documentation on moving issues: https://help.github.com/en/github/managing-your-work-on-github/transferring-an-issue-to-another-repository
<alejandra> it is simple to do
Future Work
PWinstanley: I have a list of what we
are going to work on
PWinstanley: are people happy with what's been tagged as future work?
roba: shouldn't this be a dcat subgroup discussion
SimonCox: the subgroup's not meeting anymore.
PWinstanley: is it going in the right direction?
riccardoAlbertoni: I agree with the direction we are taking. It's important that we finish the things we set aside to move to recommendation. sooner or later we should start again with some kind of meeting to reduce the number of issues and decide which issues to pick up with first.
<AndreaPerego> +1
Conneg
ncar: we had a mention a couple weeks ago and said on the horizon was the completion of a test suite. That is close. It's running, and we've tested a couple instances.
PWinstanley: is this with QSA or header?
ncar: it's with headers and the way we present media types. In the QSA one, we have the opportunity to use URIs. We've fallen back to string for everything, so the toolkits are consistent there.
… The main outcome is that in the document there are a series of requirements that are not themselves in the document.
PWinstanley: is this things like delivering profiles that are not available?
ncar: more fundamental than that. Every response has to tell you what profile it conforms to, and then more stuff. We're creating an appendix that gives the details.
roba: That will simplify reporting, not just testing
PWinstanley: what stage are we in with the document?
ncar: it was frozen at PWD3 we wanted to do testing and see what comes out of it.
ncar: we don't see any required doc changes
roba: there is one, how tokens in the alternative QSA model are expressed.
ncar: now that we've done testing, we should just itemize the requirements better.
PWinstanley: the goal is to do PWD3, put in the appendix, then go for CR?
ncar: yes, my guess is that we are pretty much ready to go. should that be before or after the repo split?
… if the group goes with splitting, I'm happy to make that happen.
PWinstanley: I think we are, based on the decision above.
roba: If you're splitting off, you're not dependent. It's just the cleaning up of issues. I think we're pretty safe.
PWinstanley: We don't want to be running two repos concurrently.
… we need to avoid putting stuff into the old one once the new one is there.
ncar: I'll wait until there's a clear "here's the repo"
PWinstanley: that's something we need to talk to Phillipe about, isn't it?
ncar: they can easily spin up a new w3c repo.
PWinstanley: we need to be applying the rules, checking they have it on backup/archiving.
Prof
ncar: we have generated one or two new issues, but the feedback we received was that we might consider supproperties of profileof/isprofileof. This could be an extension or in the vocab proper.
Rob and I have done some profile recursive validation using SHACL files.
roba: I'm going through some implementation things. I've got the conneg mechanism running, still working through the governance procsess of how profiles are published.
PWinstanley: thanks all!
<ncar> We (Rob & I) are traversing profile heirarchies like this recursively for validation file s(SHACL): https://www.w3.org/TR/dx-prof/#eg-hierarchy
<alejandra> that's nice ncar
<riccardoAlbertoni> Thanks a lot! Bye ..
PWinstanley: we need to engage with Phillipe. Let's make sure we use the mailing list and the existing repo for issues for now.
<AndreaPerego> Thanks, bye!
<Ana> thanks all, bye!
<PWinstanley> bye
<alejandra> bhe!
<alejandra> bye!
<roba> bye