Meeting minutes
<gkellogg> https://
<gkellogg> https://
Introductions
ora: Ora Lassila, Amazon Neptune DB team
… I was there in 1996, co-author of the original RDF spec
pchampin: W3C fellow, staff contact of the WG
ktk: CEO of Zazuko, co-chair of the WG
AndyS: Apache Software Foundation, been working on RDF and SPARQL for years
gkellogg: perma-invited expert in RDF related WGs (JSON-LD, RDFa...)
… have implemented the RDF stack in Ruby
Souri: Oracle, working on graph data; been involved in previous groups (SPARQL), editor of R2RML
AZ: associate prof. in École des Mines de St Étienne; discovered RDF in 2004 and have worked on it ever sinced
… was active in the OWL2 WG
EFranconi: prof at university of Bolsano; research in knowledge representation, description logic
<AZ> AZ: was also active in RDF 1.1 WG
EFranconi: have contributed to RDB2RDF
… interested in KR and reification
dchaves: post-doc researcher in @@ Belgium and in Madrid; chair of the CG on Knowledge Graph construction
Dominik_T: work at university; work with RDF and Property Graphs, involved in the LDBC Property Graph Schema WG
… that's why I'm interested in RDF-star
Doerthe: post-doc at TU Dresden, interested in Notation 3 logic, and how it relates to RDF-star
gtw: Amazon, and also interested through open-source projects in which I'm involved
olaf: senior ass. prof. in Lkinöping university (Sweden) and Amazon scholar with Ora and Greg
… author of the original RDF* proposal; have worked in the PROV WG
rubensworks: Gent University, work on JS implementations of RDF / SPARQL
… interested in implementation and query processing
remiceres: Wimmics steam at Inria; work on Corese
… will implement RDF-star in the future version of Corese
TallTed: OpenLink Software; we make Virtuoso and other RDF/LD tools
Ora: it is great to have so much expertise about RDF and co. in this group
What we expect from that group
ora: in the Neptune team, we have a model called 1Graph, unifying RDF and PG
… Neptune gives users the choice between RDF and PG, which people find confusing
… they don't know what to chose, and that hinders adoption of graph technologies
… RDF-star is seen as an important part of this work
… We are also very interested in supporting RDF-star
EFranconi: I'm here to listen to what are the real requirements for RDF-star
<Doerthe> @pchampin you put Dominik's name on my data in the protocol (just for the record, I am sure that can be fixed afterwards :) )
EFranconi: Want to check that the choices scale up to other languages (OWL, etc)
… [link to relational databases and conceptual modeling]
AndyS: my interest comes from the reification side of things
… reification has not been widely accepted, for various reasons
… we need to address these issues
… I think it is useful to have the ability to talk about triples for provenance and such -- one of the use-cases of RDF-star
AZ: like Enrico, my interest is to see how RDF-star semantics interact with OWL and RDFS semantics
… also interested in contextual reasoning, where RDF-star could be a candidate
ora: to react to what Enrico said; I always considered reification as one of the most misunderstood parts of RDF
… Early, I pointed out that reification was going to lead us in some sort of modal logic,
… as soon as I did, nobody wanted to touch reification. That's why reification ended up without any formal semantics.
… An alternative to having a formal semantics is what I call "representation by convention".
ktk: came to RDF as an engineer because it solved problems I had.
… often talk to people that I need to convince that RDF is useful compared to PGs.
… It would be good if RDF-star could stop those two communities fighting.
… RDF-star could also help in the domain of data governance.
Souri: I implement the standards developed at W3C. I am also involved in PGs.
… Like Neptune we support both RDF and PGs. We want interoperability to be as seamless as can be.
… I see two needs: 1) being able to say something about a triple
… 2) being able to have multiple occurrences of an edge
… e.g. me working for a company during two different periods in time
… RDF-star does not support the 2nd point.
… Implementing standards take time, so the standard should not go halfway.
… I suggest that RDF-star not only supports "quoted triples" but also "named triples".
… Naming is not always necessary, but would be an option.
<Zakim> rubensworks, you wanted to discuss RDFa
<rubensworks> https://
rubensworks: would like to discuss the issue of RDFa
… Sarven suggested some time ago to also extend RDFa for it to support RDF-star
Dominik_T: interested in serializations. RDF has standard serializations, but PGs don't.
… Could RDF-star be used to provide PGs with a standard serialization.
<Zakim> gkellogg, you wanted to ask if named graphs might be a surrogate for named triples.
AndyS: glad that we get to update the specs, that need to evolve (a number of errata)
gkellogg: to respond to Souri's request for considering of "named triples"
… seems related to the unfinished story of named graphs, that do not have a formal semantics
… named graphs containing triples can be considered to be different things, if we figure out their semantics
… This is done in practice; JSON-LD makes it possible to create anonymous named graphs, and VC uses this to talk about the graph.
<csarven> Issue ( Include RDFa Core 1.2 as a normative specification to deliver): https://
gkellogg: About RDFa, there is a large technical dept on RDFa (no support for named graphs).
Souri: I agree that we have some issues with named graphs. What do we do with inferences across named graphs?
… But naming graphs is different from naming triples. A named graphs is a *collection* of triples.
… The "singleton graph" idea was suggested earlier, but customers are already using named graphs for their own purposes.
… This would require them to change their practice.
ora: we see customers using named graphs as a replacement for reification; I'm not terribly happy with that.
… Souri: you have a paper about the notion of "named triples", right?
Souri: not a formal paper, but a two pages informal paper. I also can present slides.
TallTed: in my experience, people that want to talk about 1 triple find out that they actually want to talk about a set of 1 or several triples
… going along with that, we come to the notion of "collection graphs" or "graphs of graphs"
<Zakim> rubensworks, you wanted to discuss SPARQL 1.2 CG
<AndyS> Lists are one example
rubensworks: technical question: per the charter, new features in SPARQL, as suggested by the "SPARQL 1.2" CG, are out of scope. Correct?
pchampin: the RECs we will publish will not include other features than RDF-star
… but they will "allow new features" per the new W3C process, without going through the whole lifecycle
AndyS: also, different features interest different communities
rubensworks: does it mean that we have to wait until that group's work is finish to start adding new features?
AndyS: just like we did with the RDF-DEV CG for RDF-star, a CG can start working on new features in parallel with the WG
csarven: if a proposal was made to this group, with an editor's draft for, says RDFa-star, and people willing to work on it,
… would this group accept it as a new deliverable, or would it have to wait for a new WG?
… I think the expertise in this group would make it a good platform to adopt such a document. Better than creating a seperate WG.
ora: there are two things: adding RDF-star features to RDFa, and adding other things to RDF-a. The second would be out of scope.
Next meeting
ora: would the timeslot of this meeting work for everybody?
… also, would we meet once a week?
<TallTed> this time conflicts with Privacy IG and CG. Not a blocker for me, but awareness may matter...
ora: I think once a week is good to start. Maybe later we can setup task forces.
TallTed: this timeslot conflicts with the Privacy IG and CG.
… About the adoption of new work items: this can happen (that happened with JSON-LD and the RDF 1.1 WG).
AndyS: most work nowadays happen on github repositories, through issues and PRs
… I assume that we will work that way.
… Call time can be dedicated to more administrative issues.
ora: we are reaching the top of the hour.
… I propose that we reconvene next week same time. We can reconsider the schedule then.
<TallTed> +1 start at x:00, end at x:55
<csarven> pchampin: sorry, I forgot to present+. Possible to amend the minutes?