Meeting minutes
<TallTed> gotta have the RRSAgent to get the logs, minutes, etc.
Scribe: Gschwend, Adrian (alternate: Pellissier Tanon, Thomas)
approve last week's minutes
<pfps> minutes look good to me
<pchampin> Présent+
<ora> proposal: approve last week's minutes
<ora> +1
<ktk> +1
+1
<olaf> +1
<gkellogg> +1
<rubensworks> +1
<doerthe> +1
<AndyS> +1
<TallTed> +1
<gtw> +1
RESOLUTION: approve last week's minutes
Update on Use Case Proposals
pfps: It would be nice if people would contribute more use cases
pfps: the CIDOC-CRM usecase is progressing but what the user want is still unclear
pfps: We don't have many usecases but the ones I have seen have different semantics, either opaque or transparent
<pchampin> I'm interested in the wikidata UC
pfps: Wikidata has a lot things that RDF won't get. I would like to see what part of Wikidata they are the most interested in to figure out what semantic would support the usecase
pfps: This way we can take this into account and either accommodate the usecase or consciously choose to don't support this usecase
ora: anything else?
pfps: One more thing, I sent emails to a few mailing list (semantic web) to ask for more use cases but didn't get any response
<pfps> I sent out an email asking for new use cases.
pfps: I plan to build a list of all the usecases to be used as an input for the working group
doerthe: Question about transparent vs opaque semantic
pfps: opaque, transparent, and semi-opaque are three option about the the meaning of the quoted triple. opaque = what matters is the syntactic form, transparent = what matters is the meaning, semi-opaque = for blank node, meaning matters; for others, it is only the syntax. The CG followed the semi-opaque route.
pfps: Do we want to support usecases where blank node identifiers matters? I have tried to summarize neutrally the use cases on the wiki page
ktk: I see Richard Lea from Rakuten for the first time, welcome to the group. Also We got an email from George Daquila from Google. It is nice to get feedbacks from property graph people
richard_lea: Happy to join the group.
<pfps> I send out email asking for new use cases or updates on old use case
AndyS: It would be great to have definitions on opaque/transparent/semi-opaque to refer to
pfps: What about the semantic group meets and write a recap of the definitions
<gkellogg> Maybe in RDF Terminology: https://
ACTION: pfps to ensure that semantic task force writes up something on opaque vs transparent semantic
<ghurlbot> Created action #65
Update on Semantic TF
ora: Enrico is not here, should we postpone
pfps: Yes, we should
ora: Let's postpone
Review of open actions, available at 2
pchampin: Contribution from not-working group members: we will be warned if pull requests are opened
<TallTed> Only have to explicitly `close` an agenda item if zakim complains that it was just opened. Always have to `open` or `next` after an explicit `close`.
Tpt: TallTed -- Thank you
<richard_lea> I see ktk mentioned my name in a sort of regexp-like writing with slash followed with double question mark, which is sure whether a kind of predefined command in this IRC programme?
pchampin: If you go to https://
pchampin: We created all the shortcuts we decided
pchampin: This was action w3c/
pchampin: w3c/
pchampin: In w3c/
<TallTed> richard_lea — that was a command to the scribe bot, which processes https://
pchampin: Echida should work on REQ documents, let's keep w3c/
<TallTed> richard_lea — learn more here https://
<Zakim> gkellogg, you wanted to ask about Echidna
<ktk> richard_lea: can you write "present+" on the console?
<ktk> here I mean
<ktk> ok works as well :)
<richard_lea> Like this?
<ktk> richard_lea: yes, do that once you join for a call so it's recorded
gkellogg: We need a pull request to kick up Echida
<richard_lea> However I could not find the definition about the present followed with a plus symbol
<richard_lea> in that page of scribe.html
pchampin: A pull request needs to change something in the published document to get published by Echida, editing for example the README is not enough
<ktk> TallTed: it does not seem to remove me as scribe, even if I explicitly set "Scribe: ..." any ideas?
<TallTed> s/tallted:/ktk:/
Review of pull requests, available at 3
ora: Where are we with the mobile CSS?
[no real answer on mobile CSS]
gkellogg: w3c/
<rubensworks> w3c/
rubensworks: This PR is adding quoted triples to the CSV/TSV. It was not in the CG report
ora: Do you want more people to take a look
rubensworks: Not necessarily, just wanted to highlight this PR
ora: Anything else? ... Nothing, ok
ora: I urge people to take a look at PR to see if you could review/comment
pchampin: If we have a few minutes, could you briefly explain the ideas of the CSV/TSV pull request?
<TallTed> richard_lea -- Documentation of `present+` is at <https://
rubensworks: It's straightforward. It inherits the existing grammar with "<<" and recursively encodes with the existing rules for CSV and TSV
pchampin: If I understand correctly there is an ambiguity with IRIs and strings
rubensworks: Yes, indeed.
pchampin: We can live with that, CSV is already a lossy format
<TallTed> s|s/tallted:/ktk:/|s/TallTed:/ktk:/|
Discussion on named graphs 4
ora: We have an issue around named graphs. This topic keeps coming up and Dan Brickley replied to George from Google and made comments on named graph
ora: This is an issue we need to address in this group
AndyS: This is the question about the named graph and the meaning of graph name into triples
AndyS: We have the opportunity to define a new syntax. The existing solutions based on named graphs are not always very usable. It is similar to language direction.
ktk: We talked about touching named graph but we ruled it out because everyone in "street RDF" have opinion on named graphs
ktk: However, we get a lot questions about graphs, why triples stores when we manipulate often quads...
ktk: We should maybe write some best practices
ktk: When we talked about profiles, I was thinking that we could get different "profiles" on named graphs
ktk: Named graphs are hardly reusable because they get very different usecases
<pchampin> +1 ktk
ora: In my work I am using named graphs only in internal usages
pchampin: I generally agree. One think about keeping named graphs internals, one usecase where it might be hard is Solid/Linked Data Platform (LDP): a LDP platform is a set of named graphs and Solid want to build access controls.
pchampin: Named graphs are in the public space in this case
pchampin: It might be nice to get a profile for these usecases
doerthe: I agree, there are multiple semantic for named graphs, we should at least write down the relation between the RDF-star semantic we have chosen and possible named graphs semantics
ora: Might I suggest someone writes something about it to be able to point to it
doerthe: The problem for me is that RDF-star semantic is not fixed, when it's fixed I would be happy to write something
<doerthe> for the record https://
<ktk> Look what I just found googling doerthe: https://
<ktk> when was that?
<Zakim> TallTed, you wanted to remind that quads originated as subject–>predicate–>object–>*context*, and graph-name has emerged as the most common, but not the only, context
pchampin: Here is a proposal: we can point the notes from the previous working group on possible semantic of named graph. https://
triples semantic. The reason we don't started with named graphs is because they do not have proper semantic.
<doerthe> :)
<TallTed> quoted graphs are dangerous ... but feel so much more logical than limiting to quoted triples
AndyS: We will very quickly get into the "quoted graph" issue and we will get from quoted triples to quoted graphs topic. We should making clear what the boundaries of our work are.
AndyS: They were about maybe 3-4 cases on named graphs semantic in the original DAWG. The graph name is indirectly associated with the graph. It is most clearly explained in the graph store protocol schema https://
<pchampin> And there was the 'named graph' paper by Hayes and Carroll, whose proposal is very different from RDF 1.1's "named graphs" :->
Any Other Business (AOB), time permitting
ora: Time is over, regrets for next week
<TallTed> pchampin -- it'd be great if you'd comment about that difference, in w3c/