– DRAFT –
Dataset Exchange Working Group Teleconference

<kcoyle> https://‌www.w3.org/‌2018/‌01/‌16-dxwg-minutes

approval of last week's minutes

Resolved: approve last week's minutes

Resolved: approve minutes of January 16 2018

<kcoyle> https://‌docs.google.com/‌spreadsheets/‌d/‌1_f5CAZv7rgUjJH5YXkmD6w5xS6GX5X2AmEp_LMptAfQ/‌edit#gid=0

second round of UCR notices to interested people. If you have sent something again, please go to the link above and indicate that you've sent a second notice.

alejandra: I guess we are all on the email comments list, have you seen them?

kcoyle: no

alejandra: will try to find them.

kcoyle: they do show up on the comments list

<kcoyle> https://‌lists.w3.org/‌Archives/‌Public/‌public-dxwg-comments/

<kcoyle> https://‌lists.w3.org/‌Archives/‌Public/‌public-dxwg-comments/

off of our home page, comment archive

<AndreaPerego> https://‌lists.w3.org/‌Archives/‌Public/‌public-dxwg-comments/‌2018Jan/‌0000.html

kcoyle: not sure where those emails go to…

<AndreaPerego> https://‌lists.w3.org/‌Archives/‌Public/‌public-dxwg-comments/‌2018Jan/‌0001.html

kcoyle: we need to contact Dave to see if he knows who that's mailing to. Alejandra got it, but to a different email address than usual.

<AndreaPerego> I got those mails.

kcoyle: I think we should all be subscribed to that list. I will dig through and see if it just went into junk.

I didn't get them

alejandra: there was at least one issue on github tracker related to UCR.

alejandra: I'm talking about a different one. notifications on github, following the new procedure from the spatial data group for working with github.

kcoyle: Dave Raggett will look into how to implement that.

<AndreaPerego> The relevant mail from SDWIG: https://‌lists.w3.org/‌Archives/‌Public/‌public-sdwig/‌2018Jan/‌0055.html

<alejandra> this is an issue related to the UCR on github: https://‌github.com/‌w3c/‌dxwg/‌issues/‌45

kcoyle: we shoulda all keep an eye on that.

kcoyle: we have comments from Anna and one other person from Norwegain Agency for Public Management in Government.

<DaveBrowning> Also apologies for lateness...

kcoyle: the latter one lists topics they are interested in rather than comments about them.

kcoyle: we need to address substantial comments (e.g., Anna's) and reply to them.

Action: Ixchel (and Rob and Jaroslav) - look at the comments received and bring their ideas to this group before answering

<trackbot> Created ACTION-76 - (and rob and jaroslav) - look at the comments received and bring their ideas to this group before answering [on Ixchel Faniel - due 2018-01-30].

<Ixchel> Where are the comments? I haven't seen anything come through.

<AndreaPerego> https://‌lists.w3.org/‌Archives/‌Public/‌public-dxwg-comments/‌2018Jan/‌0000.html

<kcoyle> https://‌lists.w3.org/‌Archives/‌Public/‌public-dxwg-comments/‌2018Jan/‌author.html

<AndreaPerego> https://‌lists.w3.org/‌Archives/‌Public/‌public-dxwg-comments/‌2018Jan/‌0001.html

kcoyle: go to our home page and follow the link to our comments archive.

kcoyle: any questions?

silence ensues

kcoyle: if there's discussion you should post to the public list where we can talk about it.

roba: shall we discuss the issues? they all relate to scope.

roba: I"m happy to lead discussion later

kcoyle: andrea gave us some new wording for one of the use cases. We'll discuss it next week, when people have had time to look at it.

on to open action items

kcoyle: action 45 we may be able to close soon. action 50 is waiting on Jaroslav…

kcoyle: we can close action 72

kcoyle: DaveBrowning: action to contact the POE co-chair.

DaveBrowning: it will be mid-February, but there will be something coming in from there.

kcoyle: we can consider that action closed.

kcoyle: I didn't complete 74, but I did complete action 75. Simon found a great wiki page where they did a good cheat sheet for desribing RRSAgent and how it saves minutes etc.

I'll send that to the public list.

It seems the groups have all figured out how to do the agendas and minutes?

<AndreaPerego> These instructions can be included in any meeting page via macros - see, e.g., https://‌www.w3.org/‌2017/‌dxwg/‌wiki/‌Meetings:DCAT-Telecon2018.01.17

alejandra: I think it was Andrea who knew how to do that. Question: what happens if a particular subgroup has a call and noone can do it. Can it be done afterward?

kcoyle: no, but you can log it. We may need to develop our own procedures. It's not difficult.

roba: put it on the meetings page?

<AndreaPerego> The instruction are in two separate pages: https://‌www.w3.org/‌2017/‌dxwg/‌wiki/‌Zakim-DCAT & https://‌www.w3.org/‌2017/‌dxwg/‌wiki/‌Bots

kcoyle: maybe we should just copy that to the actual meetings page, so it's not buried inside a meeting.

<AndreaPerego> They can simply be loaded in any page by using {{:Zakim-DCAT}} and {{:Bots}}

<alejandra> do you mean in any wiki page?

kcoyle: what about instructions to create an agenda?

AndreaPerego: you can use the macros above to load content inside another page. We can modify the pages as we like to include Zakim and bots.

There is a different IRC channel than for normal calls for each subgroup.

<alejandra> the meeting page is created manually in the wiki, right?

<roba> * i can't believe no one guessed that syntax :-)

<PWinstanley> https://‌www.mediawiki.org/‌wiki/‌Transclusion

If you use the macros, they can be copied in automatically and updated when you update the master.

kcoyle: I'll contact you directly for clarification

AndreaPerego: in the DCAT meeting, we didn't have actions recorded, so we have to tell it to monitor the right IRC channels.

I'm not sure how to fix it.

kcoyle: it would be helpful if you could write up an email saying what you have here and this tracker problem.

AndreaPerego: I'll do that.

Action: AndreaPerego to write up a procedure for setting up subgroup meetings and to find out about the tracker issue.

<trackbot> Created ACTION-77 - Write up a procedure for setting up subgroup meetings and to find out about the tracker issue. [on Andrea Perego - due 2018-01-30].

alejandra: on a related note, I don't know how one sends an email only to members for the webex password. kcoyle, can you send it around?

kcoyle: Dave R set that up for one meeting.

alejandra: we are considering a change in time. Will the link change?

kcoyle: yes, and only Dave can make that change.

alejandra: He already made the change, but we need to distribute it to the members only list. I'm not sure what that is.

kcoyle: I thought he sent it to the list already. That is what you put in the meeting agenda.

alejandra: can I forward to you for distribution?

kcoyle: I think it needs to go into your agenda.

kcoyle: also, there is the host key. without that you don't get sound.

alejandra: last week, Simon did all this, but he's not available this week.

kcoyle: you may need to get hold of him or Dave to find out the key.

it's specific to that webex reservation.

I only have the one for plenary meetings.

kcoyle: let's move onto the task of defining profiles well enough that the negotiation group can go off and do their work.

I'm sorry that Ruben isn't here for this discussion.

<kcoyle> https://‌www.w3.org/‌2017/‌dxwg/‌wiki/‌ProfileContext

we had quite a bit of discussion on the profile page, link above.

<kcoyle> https://‌lists.w3.org/‌Archives/‌Public/‌public-dxwg-wg/‌2018Jan/‌0033.html

I added a link to an email from Lars which seemed a minimalist suggestion for requirements for identifying profiles in order to negotiate.

LarsG: for the sake of content negotiation, profiles are fairly opaque. It's mostly a matter of grabbing by URI and seeing if it's supported.

For other cases, there can be subprofiles, or ones put together from different parts.

We can think about what do we give back to the client, and how does the negoatiation work.

roba: quick question: looking at previous examples, rather than using IRIs, I tend to use short tokens that are registered somewhere. Do you intend to use IRIs or tokens or both?

<AndreaPerego> +1 from me. LarsG proposal fits with https://‌www.w3.org/‌TR/‌dcat-ucr/#ID30

LarsG: My choice would be to obviate having a registry and go with IRIs.

The threshold for creating a profile should be as low as possible.

roba: I think then we just need to update our examples.

at the top of the page.

roba: The other aspect of this is around inheritance, the duck typing problem. What subprofiles are supported, whether it's the responsibility of the client or the server.

We need to be clear about the implications.

LarsG: At this point, we don't know who is responsible. We need to figure that out as part of the work.

kcoyle: would anyone like to venture a brief definition/description of what we see as what we've defined here?

roba: looking through the comments, the main source of confusion seems to be the use of words with multiple scopes.

There's also confusion between specification and profile.

Probably these things are fairly synonymous, and we should avoid making artificial distinctions.

From the profile negotiation perspective, there needs to be a single definition of a profile. We seem to agree on what it is—a set of constraints.

The description of the set of constraints is another concept. We don't yet know how to describe these, so we probably need to defer that problem (the profile implementation specification, profile definition) until later. We need to decide if that is all one thing or it can refer to other specifications.

We seem to agree on profiles acting like objects with inheritance. That's probably the main touchpoint for what we need to define for profiles. what mechanisms exist for client or server to understand how profiles relate to one another.

<Zakim> AndreaPerego, you wanted to say that a profile, from a negotiation perspective, may not imply the definition of a set of constraints

AndreaPerego: In my understanding, from a conneg perspective, we have a use case where you have a customer providing records in different formats. For instance, the OAI-PMH protocol, used by digital libraries, can return records in Dublin Core, DataCite, etc.. I think this is an exmaple of profile negotiation. The profile is the metadata schema, and, in the case of DCAT, no constraints are specified.

AndreaPerego: probably we should relax a bit what we mean with the word profile

roba: I think unless the negotiation needs find grained understanding, we need a pretty simple thing. We don't have to negotiate on schema and profile, just profile.

the definition is as relaxed as it can/needs to be.

kcoyle: asking Rob, could you develop a proposed wording for us that we could vote on?

roba: for next week? It's not trivial.

kcoyle: we want to get the simplest definition that doesn't restrict us too much to follow our guidance.

kcoyle: I assume there will be some attempts at best practices in our guidance. Does that meet other people's view?

Action: roba bring a short definition of profile that we can "vote" on next week

<trackbot> Created ACTION-78 - Bring a short definition of profile that we can "vote" on next week [on Rob Atkinson - due 2018-01-30].

kcoyle: the profile negotiation group is going to be meeting on Fridays at 9am CET. If others are interested, contact Ruben and Lars.

roba: I'll attempt to have a strawman definition for the group on Friday.

kcoyle: so far we don't have a chair for the profiles group. Who can take this on? they need to schedule meetings, make sure webex is available, get info out, figures out how to do agendas. The difficult part is only the initial startup.

kcoyle: FPWD is due in about 10 weeks. I want to thank the conneg and dcat folks for getting stuff going for their subgroups.

I see a lot of activity happening.

roba: question about comments. Someone gave us a fairly challenging comment, use case 5.5.2, comment 01.

<roba> https://‌lists.w3.org/‌Archives/‌Public/‌public-dxwg-comments/‌2018Jan/‌0001.html

roba: anyone have any perspective on that?

do we want to make constraints deal with combinations of vocabularies?

kcoyle: I think we already have a use case that covers this

roba: I worry about the arrangement of it.

kcoyle: it depends if you're talking about RDF. I've seen this before. I'll send out an email about it.

alejandra: we could ask them to provide an example as well.

<AndreaPerego> +1

kcoyle: I think I can find some easily.

<AndreaPerego> bye

bye!