– DRAFT –
DXWG Plenary

<PWinstanley> proposed: accept minutes of https://‌www.w3.org/‌2019/‌07/‌02-dxwg-minutes

+1 approve minutes

<alejandra> +1

<PWinstanley> +1

<riccardoAlbertoni> +1

<DaveBrowning> +1

<annette_g> +1

Resolved: accept minutes of https://‌www.w3.org/‌2019/‌07/‌02-dxwg-minutes

Peter: Meeting today will continue ongoing discussion of "profile"
… Kicking off next few DCAT meetings
… And CONNEG status
… Given that we have extension - confirm that people not going to TPAC for meeting?

DCAT

Peter: Dave, Alejandra: planning over next week or so?
… Should we set up meeting with Philippe circa July 17?

Dave: Partial agenda for DCAT meeting tomorrow.
… Behind on reviewing comments. Do not know what we might

SCRIBE JUST LOST AUDIO

DaveBrowning: may need more time before meeting with philippe

<alejandra> +1 to have more time

DaveBrowning: there is work to be done before
… question for tomorrow's meeting: is what we have to concern ourselves with is the blackout period around TPAC?

PWinstanley: I'll ask Philippe about that

riccardoAlbertoni: agree that we need more time because of amount of feedback
… need to do some planning

alejandra: also agree to push meeting back
… in DCAT group we don't have a way to connect unless Simon sets up a webex
… since Simon isn't in the call do we have another option?

DaveBrowning: looks like the meeting is still scheduled on webex

alejandra: is there an agenda?

DaveBrowning: a draft agenda, yes. It's on the wiki

Profile definition and new use case

PWinstanley: lots of conversation so far. No one here today from S Hemisphere, but are there other comments?

<TomB> SCRIBE IS BACK

aisaac: I'd like to get rid of Rob's use case from the discussion

Antoine: Notion of profiles that play different roles.
… For the guidance and PROF vocabulary - centered around this notion.

Peter: You propose whether to accept the use case?

<AndreaPerego> Can we add the link to the proposed UC here?

Antoine: It needs tidying.

Annette: Just needs to have example added.
… I could not vote for it in current form.

Peter: Propose that we include pending tidying up.
… If we can agree it is something we can work on.
… Otherwise still hanging around.

Annette: Oddities about how it is written

+1 Annette - needs more clarity

<AndreaPerego> Proposed UC: https://‌github.com/‌w3c/‌dxwg/‌issues/‌978

Karen: More sceptical. Tautology here: specs that have certain qualities in them.
… Am pretty sure there are existing profiles that could not be described this way.
… If not based on particular standard... A particular style of profiles.

Antoine: Maybe does not capture everything, but captures several cases we want to answer.
… Terminological issues - essential problems around "specification roles" - needs to be cleaned.
… Ready to accept that more work needed - more than "tidying up".

Karen: Someone will pick an example that meets this use case, but not examples that do not.
… One example is not enough - need broader view. Not universal enough to fit in our work. More like a long-tail sort of thing. Relevant for specific communities.

Antoine: Agree in theory - alot of what UC focuses on should not happen. We know this is not always be the case. Relevant to definition of specific component. No one-to-one re: roles.
… Would hope we would not have to address the case of representation - confusing in real world - good to be ready. Would prefer Rob to elaborate.

<kcoyle> "Documents that perform multiple roles, such as multiple discrete sets of requirements, suggestions etc need provide stable identifiers for each logical grouping of specifications."

Andrea: Unsure about this UC. What is missing is problem statement. There was some discussion on GH about multiple roles. But issue that derives from our view of profiles - that they have "roles"
… But reader of UC document might have no clue what the problem is.
… Difficult for people not involved in this discussion.

Karen: Posted (above): "documents that perform multiple roles" - and we all recognize that there may - "discrete sets of requirements" - "each logical grouping of specifications".
… It is not a database but a specification document.
… I don't like calling them roles.
… The part about identifying different parts makes no sense.
… Fine if you have grouped things this way, but what about DCAT? Must address.

Andrea: Fact that we have multiple roles - a matter of fact. Why do we need a UC? Or can we simply state it?

<PWinstanley> TomB: The more I hear about this UC the less I like it

<PWinstanley> ... it sounds to me like - I don't like the idea of looking at things as roles, and looking at a profile as though it has discrete sets of anything

<PWinstanley> ... I know that machine-ready profiles would have units

<PWinstanley> ... are we talking about what profiles *are* or on what they should be?

<PWinstanley> ... This WG has yet to take a decision on whether to continue work on the profiles vocab. There are requirements for DCAT and CONNEG on what profiles *are* but the vocab is looking at what profiles *should be* - and we might not have the time for this

Antoine: Re: what profiles should be - but we can accept a use case - hint that it is a possibility.

Annette: This is helpful if we think not of single documents that break into parts, but collection of expressions of profiles.
… Someone might publish guidance: machine-readable + plain text + images...
… When you have a collection of guidance along with machine-readable schema.
… Feel like rather than saying "profiles" negotation, but "schema" negotiation - the machine-readable part.
… We need to sort this out because it is important. Agreed definition that works would be a big contribution.

<aisaac> annette_g said it's not only about the machine-readable part (I really want this to appear as it's important ;-) )

Peter: Caught up in words? Venn diagram approach better?

Annette: Everyone has pieces that they are important.

Peter: Enterprise Architecture has "viewpoints" - define specific VPs?

Annette: Approach this afresh. Agree on concepts before we agree on wording.

Karen: Can we put action on Rob or someone to make this profile-specific, not specification-specific.
… If it defines profiles, needs to be about profiles.

<kcoyle> so then it may be out of scope

Antoine: Reluctant to ask Rob to do this. Understand need for generality, but if profiles based on other things - but Rob suggesting something more general.
… If we ask him to focus on profile. This is not about things based on other things.

Karen: If it is not about profiles, it is out of scope. We are not about defining specifications. Could be useful to help clarify, but.

Antoine: Rob's use case applies to things that are [].

Karen: You say it applies to CONNEG but not DCAT?

Antoine: Depends on what we call profiles. GUIDANCE could have been about both points.
… I see Rob's UC applying to CONNEG.

Peter: But only to the bit that applies to specifications about specifications - not atomic specs?

Antoine: No, because CONNEG does.

Karen: Labels of the use case are about "profiles".

Antoine: Then we need to clarify. But in description of UC, talks about things about other things.

AndreaPerego: Sometimes we go around and around. When I see title of UC, assume it is about profiles.
… Still confused: not clear what we had decided about profiles. Just semantics, not syntax? As long as we do not clarify at this level, discussions never ending.

Peter: Talking in circles for half an hour.
… We are still pretty confused about profiles.
… I think we can park this. We have not yet decided what to do about PROF vocabulary.

<kcoyle> +1 to TomB

<aisaac> +1 the problem is that we can't agree on words...

<PWinstanley> proposed: to defer discussion of they UC and focus on definitions of profile as the *are now* for the conneg and dcat docs

<kcoyle> +1

Peter: Propose to focus on definition of profile as they are now for CONNEG and DCAT (agreeing with Tom's point).

+1

<riccardoAlbertoni> +1

<PWinstanley> +1

<AndreaPerego> +1

<DaveBrowning> +1

<aisaac> +1 it's a pity but if it can help the discussion

<annette_g> +1 for now

Peter: we need to recognize that time is running - we need to complete CONNEG and DCAT
… If we do that, will be in position to decide whether we need a Note - would be a pity not to nail it after all the work.

Resolved: to defer discussion of the UC and focus on definitions of profile as the *are now* for the conneg and dcat docs

Antoine: Very much recognize problem. If we do not give it space, will resurface.

Peter: Clarify part w.r.t. roles.

Antoine: If we can be strict about not considering "roles"...
… PROPOSE that we use Rob's UC to discuss what roles can be played with respect to specifications, and park issues there.

<aisaac> proposed: that use case #978 shall be used for discussions on the roles played by profile "representations" and that discussions about these matters should not interfere in other discussion spaces

<AndreaPerego> +1

<kcoyle> -1 if we call them "roles"

<PWinstanley> TomB: Are you saying that the discussion about the UC will continue, but on a separate track?

<PWinstanley> aisaac: yes

<PWinstanley> +1

+1 that the discussion be on a different track from defining "profile" for CONNEG and DCAT

<riccardoAlbertoni> +1

<DaveBrowning> +1

<annette_g> +1

<aisaac> +1

Karen: I really need something more specific than "roles".

<PWinstanley> TomB: I also don't like the term 'role' but agreed with the essence of the proposal that it moves to a track that is separate from the conneg and dcat discussion of profiles

Karen: I would like to see the resolution change to say...
… Rob does not use the word "role" here, in that sense. He talks about requirements - requirements embedded in documents.
… I would like to see us talk about something other than "roles" - "functionalities"? Documents do not have roles - as a desirable.

Peter: You agree that this is only partly parked, and we can pull parts of out as needed.

Karen: Would like to have it wholly parked. I'm all for parking.

Peter: So where are we w.r.t. to Antoine's - pick at it if helpful for discussion - rather than totally parking.

Karen: This idea that it is going to be used for discussions on "roles" - unclear how they fit into CONNEG and DCAT.

Alejandra: [ scribe did not catch ]

<kcoyle> alejandra said we maybe should have waited for Rob to discuss this

Peter: We are clear on finishing requirements for CONNEG and DCAT - think we need to reject proposal until we have another discussion.
… Antoine?

<riccardoAlbertoni> can we vote it saying we will change the word role

Antoine: Maybe something more precise next week?

Karen: Alejandra is right - we need to clarify with Rob how UC fits with CONNEG and DCAT.

Peter: We are parking this until clarify rel to CONNEG/DCAT

+1

<AndreaPerego> Can we have a proposal?

<kcoyle> +1 park until we clarify vis a vis conneg and dcat

Peter: We just park the resolution. Thank you all for a very engaging conversation.
… It is bringing us closer to what we need for CONNEG and DCAT.
… Hard to see wood for the trees, but at some point we will have to deliver.

<Zakim> TomB, you wanted to ask when CONNEG will come up for its next transition

Peter: Will need to find out next week.

<riccardoAlbertoni> Thanks, bye

<annette_g> thanks all!

<PWinstanley> bye!

Peter: Next week, same time, thank you all for very active conversation.

<AndreaPerego> Thanks, and bye