The W3C JSON-LD Community Group
Go Back
JSON-LD CG
Minutes for 2024-09-18
- Present
- Gregg Kellogg, Anatoly Scherbakov, Benjamin Young, Pierre-Antoine Champin
- Chair(s)
- Gregg Kellogg
- Scribe(s)
- Gregg Kellogg, Benjamin Young
- Agenda
- https://www.w3.org/events/meetings/08563546-30fd-497a-ac16-77f96e96e8c0/20240918T120000/
- Topics
Gregg Kellogg is scribing.
Topic: TPAC Preparation
Pierre-Antoine Champin: Will be a little late, sorry
Benjamin Young: I think we have the slides we need for the BOF.
... [shows slide deck]
... This presentation focuses on CBOR-LD and YAML-LD along with language patterns.
... Has bar codes based on JSON-LD and CBOR-LD to show the dramatic size difference.
... Needs YAML-LD content from anatoly-scherbakov
Anatoly Scherbakov: I haven't had a chance to look at the slide repository. I'll get them soon.
Benjamin Young: Our presentation time is 14:00 PDT.
... On Wednesday the 25th.
... We have joint sessions Monday, Tuesday, and Thursday as well as a dedicated session on Thursday.
... Not sure what the agenda for the Spacial Web meeting is.
... You can present YAML-LD on Thursday, but I was hoping to have them for the Wednesday meeting.
Benjamin Young: https://www.w3.org/2024/09/TPAC/breakouts.html#grid
... Subscribe to the event, and it will add it to your calendar.
... TPAC registration deals with regular WG meetings, the breakout needs to be added separately.
Anatoly Scherbakov: That time will be 03:00 my time, but I want to join the meeting.
Benjamin Young: My expectation is that the audience either knows or are interested in JSON-LD.
... We'll discuss CBOR in both the breakout and group session.
Anatoly Scherbakov: So, you're preparing two presentations: one for the breakout and the other for the group meeting.
Benjamin Young: Focus of group meeting will be on re-chartering.
... what we need to do to transition, and finalize for a vote.
... With remaining time, we can look into remaining issues.
... Left over joint meeting time may be hard to redirect to focused group issues.
Anatoly Scherbakov: I'll get a PR for the slides done in the next couple of days.
Benjamin Young: Feel free to just email them into me.
... Breakout should have just an example of YAML-LD.
Anatoly Scherbakov: Do we expect that OWL might be discussed at TPAC?
Pierre-Antoine Champin: OWLs are not what they seem
Gregg Kellogg: Possibly some discussion of OWL in RDF-star meetings.
Topic: JSON-LD-star presentation https://json-ld.github.io/w3c-tpac-2024-presentations/json-ld-star/
Gregg Kellogg: Discusses JSON-LD-star
Anatoly Scherbakov: What is the difference between JSON-LD reified reified station and normal RDF.
Benjamin Young is scribing.
Gregg Kellogg: Reifiers let you talk about triples without them having to be part of the data model
... they can also be used with named graphs
... the form looks similar, but the semantics are differen
Anatoly Scherbakov: I think I have read that reifiers are a problem because you can't reason on them
... so what's the point of defining them?
Gregg Kellogg: Reified statements in RDF 1.1 are different than in RDF 1.2
... in RDF 1.1 they were really only for RDF/XML
... in RDF 1.2 they are more atomic and they are resources in themselves
... they are useful
Ted Thibodeau Jr.: In the classical sense, where reification is done by declaring subject, predicate, object on an identified Thing
... it's overly verbose and painful to do
... RDF 1.2 (fka RDF-star) you can write the triple
... and simultaneously have it be in the graph and annotate it, or have it alongside the graph and annotate it.
Anatoly Scherbakov: Yeah, I've wondered how easy this will be to explain
... even if you have named graphs it might happen that you have them because they came from different sources
... and it's easy to get into problems when you use an OWL reasoner
... because they all come from different sources and may conflict
... do you think that will be solved in RDF 1.2?
Ted Thibodeau Jr.: It may be solved over time.
... right now it comes down to your processor and SPARQL store
... and how you scope your query
... it's definitely a problem that comes up
... depending on how you populate your store
... We have a thing called the Spounger
... it populates a graph named for the source (and possibly the time stamp when you consumed them)
... things change over time, so that's important also
Pierre-Antoine Champin: To compliment...what's an OWL reasoner to do?
... it doesn't say anything about datasets
... when datasets were added to RDF 1.1 they had been around in SPARQL for awhile
... they'd been used for all kinds of things
... so standardizing it then would have broken other use cases
... the name may be used for different things
... quoted triples (aka triple terms) we can now define proper semantics for that
... SPARQL and OWL would be updated via different WGs which don't exist yet
Anatoly Scherbakov: Thanks. I'd love to see what's happening there and maybe take part
... but I have minimal understanding of descriptions logic on which OWL is based
... but I'd love to help as I like RDFS and OWL
Gregg Kellogg: There is a hope SPARQL, OWL, etc. will be updated, but we will need a WG
... also tools like SHACL may get a new WG
... and it would use triple terms
... there is a charter circling for that
Anatoly Scherbakov: But SHACL is validating a shape
... OWL doesn't care, it'll infer anything it can
Ted Thibodeau Jr.: Right. different, but they connect
Gregg Kellogg: OWL can be used for cardinality restrictions
Pierre-Antoine Champin: I will have to disagree, gkellogg
... cardinality constraints aren't helpful with the shape of the graph
... if you say one must only have 2 parents, and then use OWL to say they can only have 1
... doesn't do what you want
... it can be used for validity constraints...a bit
... SHACL actually can be used as an alternative to OWL for reasoning
... one work item in the proposed SHACL WG is SHACL Rules
... so it's starting to grow beyond shape validation
... the shape can be used at the head of the rule
Gregg Kellogg: Is that a SPIN derivative?
Pierre-Antoine Champin: I don't know
... there have been some experiments, though, so it's going into the WG recharter.
Gregg Kellogg: Yeah. OWL is basically too obscure now
... and SHACL is kind of taking over
Anatoly Scherbakov: Yes, OWL is hard to debug...and needs tools
... at the modern level of modern IDEs
... and it's had it's own controversies
... like how much of the graph one imports when reasoning.