Meeting minutes
Announcements and newcomers
Dominik: my first time here. I work in University in Poland. subjects RDF and Property Graphs.
… I'm also in the N3 CG, LDBC Schema WG also working on Property Graphs and Schema.
<ora> RDF and PGs, sounds interesting!
Open actions
<pchampin> https://
pchampin: Some of these actions were about reaching out to developers.
ora: I was to write a blog post and to talk to PatH.
pchampin: I ment Olaf :)
<pchampin> https://
olaf: I emailed Ontotext and Pavel from Star Dog. They both responded.
… Pavel has to check with the rest of the company, and wants to look again at our report.
… In which case many tests may have failed and they may not be excited about publishing.
… What they had implemented was for PG mode, so even we have SA mode, most likely many of their tests will fail.
… I need to follow up with Pavel.
… Ontotext immediately replied that they would be happy to send an implementation report.
… Generally, they're positive and we can expect something sometime.
… OTOH, he also involved RDF4J as they have a shared implementation, and perhaps it makes sense to have a shared report.
<pchampin> https://
ora: I started on the blog post but got side-tracked by work responsibilities.
… It will be a blog post on the AWS blog.
<pchampin> https://
ora: I also reached out to Pat Hayes (wonderful credentials).
… We talked about the Neptune 1G effort to unify RDF and PG.
… I explained my take on RDF-star, which I considered to be important, which he understood.
… I said my worry is that we don't open the flood-gates for all kinds of changes to RDF and keep it tightly scoped.
… THen we ended up speaking about his B-Logic proposal. It was an ISWC Keynote in 2009.
<pchampin> Pat Haye's BLogic
pchampin: This pops up regularly, paerticularly in N3 CG
ora: For due diligence, we should read his slides carefully.
… THat said, I want to be sure we keep RDF-star tightly scoped.
… It will be very interesting for many people to read his ideas.
<pchampin> Pat Haye's BLogic talk
<pchampin> that is, if you have a "modern browser" with flash enabled!!
ora: It extends RDF semantics, without replacing any.
… It basically extends RDF to a full 1st order Predicate Calculus.
… It also has an interesting take on Named Graphs.
pchampin: I also posted a version that requires Flash to view.
… The slides are on slideshare, but there was a video lecture where you need flash.
… I agree that B-Logic is interesting, although it goes beyond our tight scope.
… I have some thoughts on defining RDF-star on top of B-Logic.
ora: Interesting is what he calls "surfaces", both negative and positive and neutral.
… This allows you to reason over things that you might not believe in.
<pchampin> https://
pchampin: This action was on my reaching out to Corese implementers. They have an implementation which complies (mostly).
… THey're willing to upgrade Corese and implement a submission report.
<pchampin> https://
pchampin: Next is RDF4J.
AndyS: I emailed Jerven, who isn't that keen on submitting a report.
… Theirs tracks more the PG mode of the old spec.
… I haven't pushed them any further, even if it is via reification, I don't know why they couldn't pass the test suite.
pchampin: The semantics might pose a problem for them.
AndyS: I believe they do have an EARL report generator.
<pchampin> https://
pchampin: The last issue was to create a draft blog post for the CG blog.
<pchampin> https://
pchampin: From my perspective, the post is in good shape. Compared with the initial plan, the scope of the post is smaller.
… I thought there was enough content. It focuses on Provenance.
… Including simple statements, and more going on to complex.
… One point was to have some arguments to bring to the issue raised by PFPS a couple of weeks ago about the examples in the report being broken because they don't use intermediary nodes.
… We worked on the Google Doc and not on the mailing list to try to constrain the conversation before it is posted.
… I propose we publish it right now.
<rivettp> looking at it now, I'd liek to chiem in
<rivettp> give me until end of day please
<pchampin> PROPOSAL: publish the CG blog post ASAP
<fabio_vitali> can I have access please? Fvitali@gmail.com
AndyS: I think we should publish it. It's not designed to be a technical document.
+1
<pchampin> +1
olaf: I didn't have a chance to look at it, but if you're fine with it I say go ahead and publish.
<AndyS> +1
<olaf> +0
<rivettp> are we using US English or British English?
<AndyS> English English
<rivettp> in US it's modeling
<rivettp> I'd vote for publishing Monday
fabio_vitali: What kind of feedback are you looking for?
pchampin: The idea was to write it collectively.
… On the blog we would credit the RDF-star task force, so everyone has a chance to contribute.
fabio_vitali: If I have some opinions on the appropriateness would you like comments in the doc, a mail, or what?
pchampin: We can start the conversation here.
fabio_vitali: One problem I have is about the negative example.
… It may be too early to talk about limitations of RDF-star. I also think there's a more correct solution to prevent the errors from occuring.
… The problem may then disappear which could become advice. (using two levels of nesting).
pchampin: I'm not sure I agree in this situation. The goal isn't to claim that the proposed solution is the only way to do it, but to highlight the use of additional nodes.
… A agree that in some cases double-nesting might be a solution.
… Do we agree that the negative example is broken?
… It's not just about bad modeling, it's "lossy".
<rivettp> OK I've finished it now and am happy with it - I suggested some changes e.g. to use "SPARQL-star"
fabio_vitali: I wouldn't say "broken", but yes.
<fabio_vitali> Sorry about the Latin, but this is a case of "Excutatio non petita, accusatio manifesta": if you make excuses that are not requested, you are accusing yourself of something nobody would have cared about.
pchampin: Lets continue discussion into the beginning of next week.
<pchampin> PROPOSAL: publish the CG blog post in the beginning of next week
<pchampin> +1
+1
<AndyS> +1
<rivettp> +1
<olaf> +1
<ora> 1+
<fabio_vitali> +1
<ora> +1
<Dominik> +1
Resolution: publish the CG blog post in the beginning of next week
pchampin: We'll continue the conversation on the Google Doc.
olaf: Since you mention DanBri, I think the final report isn't published on the CG page yet.
pchampin: Yes, I don't think Dan responded to my email.
Action: ping Dan for publishing the final report
<fabio_vitali> I added a comment on the Google Docs message
AndyS: there is a proposal from Oracle for RDFn that addresses some of the PG issues.
<AndyS> https://
AndyS: It's been around for a while, but not sure of its status.
<AndyS> https://
AndyS: It's written more from the PG outlook.
… I don't suggest we do anything about it in particular, except that we've noticed it. I don't think its anything more than a theoretical proposal.
pchampin: Should we mention it?
AndyS: Unless we do a comprehensive survey, we would have missed other things likelly.
<fabio_vitali> I have been trying to suggest that rdf-star start thinking about non-asserted named graphs, too
pchampin: Named graphs are introduced, but not the main point.
… It depends on how a property is defined, but ends up with repeated triples.
pchampin: This could be modeled or emulated on RDF-star, as it gives you the ability to refer to a triple.
AndyS: It may be a slightly higher-level model, but it has details to be figured out.
ora: I read Oracle's proposal and found it interesting. I think we should encourage them to join the WG.
… I can reach out to the author.
AndyS: I told them there's going to be a WG.
ora: We found in original RDF group that the more people you bring in the room, the better.
olaf: I wanted to say that he'll probably show up in the WG with this proposal as a counter-proposal.
… I read it some time ago, but didn't think it really did enough and doesn't have clear definitions.
pchampin: I think it makes sense to be pro-active.
fabio_vitali: This article is two years old. I wonder if they've gone on to do something with it, or it just was an arbitrary statement.
AndyS: It was last updated in 2021-09.
<pchampin> https://
pchampin: I created two small PRs to the charter about RDF/XML and additional specifications to be included in the charter.
<rivettp> regarding implementation reports has anyone reached out to Cambridge Semantics for Anzo? They have a presentation about rdf-star on their website
pchampin: This will depend on proposals emerging, and then re-charter to include.
… But, we keep the door open for producing other normative documents.
… We'll focus on the charter in the next call. In the mean time, I'll try to iron out the missing details, but please comment on the issues and PRs.
… I think we need to list all the SPARQL documents so that they can be consistent.
… I'd like us to discuss the charter and consider submitting it to the semweb mailing list before going to the official process.
<fabio_vitali> why "next week"? Not in two weeks?
<fabio_vitali> thanks
pchampin: We also need to find chairs. Ora said he'd considered (IIRC).
<ora> I would consider, depending on my employer's opinion on this.
AndyS: How many chairs are typical now?
pchampin: Now two chairs are most common.
pchampin: We also need to consider editors, but Chairs are most important and need to be on the charter.
… PhilA is probably going to chair the RDF C14N WG.
rivettp: I mentioned Cambridge Semantics.
pchampin: No one has reached out that I know of.
AndyS: I might have a contact.
olaf: I have contacts too, I'll reach out.
Action: olaf to reach out to Cambridge Semantics (Anzo)
<fabio_vitali> thank you and bye
pchampin: back in two weeks.