The W3C JSON-LD Community Group

Go Back

W3C Logo


Minutes for 2023-04-19

Gregg Kellogg is scribing.

Topic: Announcements and Introductions

Gregg Kellogg: Call for membership of JSON-LD WG
Anatoly Scherbakov: I'd be interested in being there.

Topic: YAML-LD

Anatoly-scherbakov_: The work on YAML-LD isn't yet complete.
... I've added a non-normative section comparing JSON and YAML by their feature sets, showing that YAML has some extra features.
... I started on another section discussing extended YAML-LD.
... I describe the goal of using those features for describing nodes.
... I've created some issues, one completed and merged. ANother is in a draft state.
... There is also a note on examples.
... The examples of YAML-LD in the spec should be universally themed, meaning that they should each address a different topic.
... For example, an example of my cat with names in different languages, others talk about different people.
... It would be more cohesive if they all related to a single topic.
... The most straightforward choice might be to have examples about the specification itself.
... This could be written in different variations, say using anchors.
... There hasn't been discussion about it yet.

Subtopic: json-ld/yaml-ld#94 -> Issue 94 Unify YAML-LD & JSON-LD examples (anatoly-scherbakov) : Unify YAML-LD & JSON-LD examples
Anatoly-scherbakov_: Also, thanks to TallTed for his help.

Subtopic: json-ld/yaml-ld#88

Anatoly-scherbakov_: Maybe we should reiterate the general plan: That's described in json-ld/yaml-ld#88, as a diagram. -> Issue 88 Leave YAML-LD Extended profile out of bounds of the Community Report (anatoly-scherbakov) : Leave YAML-LD Extended profile out of bounds of the Community Report
... Right now I'm documenting motivations for the extended profile.
... What's next is to complete work on the extended profile.
... Also, there are tasks on the extended test suite, where we can document downgrade tests using the JSON profile.
... My progress has been slow, I'm afraid, as I've also been occupied with experiments.
David I. Lehn: I like the diagram. Are there implementations that implement these different features.
Anatoly-scherbakov_: Yes, the diagram was taken from YAML-LD.
... I haven't been using any extended profile features, just the JSON profile using our convenience context.

Subtopic: Other YAML-LD issues

Gregg Kellogg: There are some problems with expressing IRIs containing a '#' in many YAML parsers.
Anatoly-scherbakov_: I haven't looked further into it. I think the syntax for tags in YAML is rather convoluted.
... YAML feels like a concise language, but the tag definitions do not.
... To assign types, I've either used the `$type1 property, or rely on OWL to assign types based on the domain/range.
... I haven't felt a pressing need to use the tag syntax.
Niklas Lindström: +1 For looking for pressing practical needs
Niklas Lindström: I had a discussion about YAML-LD when you stick to the "good parts".
... It would be beneficially for YAML-LD to stay within that subset.
... I'd recommend that we keep YAML-LD to using the well-implemented subset commonly implemented.
Anatoly-scherbakov_: My intention for using YAML-LD was to enable non-specialists (who may have not read the spec) to write YAML-LD easily.
... It can be effective as a DSL; they want something very simple.
... That's the motivation of moving the extended profile out of the spec, and move it into a non-normative section to fuel further discussion.
... That way, the experimental features may be pushed forward at some point in the future.
Gregg Kellogg: Our goal is to get the spec into a position where we present the spec to the WG. Then, we should get this onto Recommendation track later this year.
Gregg Kellogg: Let's push the thing forward and get it to the state of a final report.

Topic: Open Discussion

Subtopic: Errata/updates to JSON-LD specs

... JSON-LD framing has an issue that's affecting us.
... It's a fairly small change about not sure what to do about this one.
Anatoly-scherbakov_: I want to ask for a consultation, but let's come back to that.
David I. Lehn: There was a test added to the test suite, but it needs more coverage.
Gregg Kellogg: I believe I implement the test.
Pierre-Antoine Champin: About updating the recommendation; once the WG is active we can probably make some changes for that.

Subtopic: YAML-LD use cases

Anatoly-scherbakov_: I'm developing a YAML-LD application for some time. My concern is that I might be doing something that is not generally interesting to people.
... How can I ask for feedback on this?
Anatoly-scherbakov_: I'll create an issue on the YAML-LD repo as a use case and solicit feedback.
Niklas Lindström: I was going to ask about open issues for the JSON-LD WG. Should we update it, or is there an an erratum.
... I'd need to set aside some real work time to be able to devote time on this.

Subtopic: RDF-star

Pierre-Antoine Champin: At the moment, neither the RDF-star nor JSON-LD are chartered to do JSON-LD star
... This doesn't mean that it couldn't happen in those working groups, but there may need to be some re-chartering.
Niklas Lindström: I'm interested in the JSON-LD star and would have time to work on it actively.
... There are a number of leads cropping up for atomic provenance.
... I've thought of joining RDF-star as well, but was worried it would take up all my time.
... There are many interesting things happening in our data when trying to annotate provenance and starting to use that elsewhere.
... I'm glad niklas has time to work on it, and would like to collaborate.
... The more active work there is on the spec, the more chance there is that it would not be objected to.
... It would be best to continue to incubate in the CG to make it acceptable.
... The semantics discussion is taking a lot of time, but it's also quite important.
Niklas Lindström: It's hard to say away, but I'm resisting :)
... When we work with the JSON-LD star we use it as JSON. It's sort of like programming in "C", it's dangerous.
... That gets aggravated if you try to discuss provenance without controlled semantics.
Gregg Kellogg: Status of JSON-LD-star (& ⇒ YAML-LD-star) is fairly stable. Depends on what happens with Annotations. There's not much work left for the spec.
Gregg Kellogg: Feel free to take part in discussions, issues, mailing lists.
Pierre-Antoine Champin: +1
Gregg Kellogg: Sounds like there are use cases to contribute. There're ways to contribute without diving in too deep. Your contributions have been quite useful in the past

Subtopic: YAML-LD deadlines

Anatoly-scherbakov_: I wanted to ask about any deadlines, or calendar for prepariing the YAML-LD spec for the WG.
Gregg Kellogg: None, but we should capture the momentum. Once we've wrapped up the predominant issues, maybe just defer them, — and come up with an MVP useful spec. Then the CG votes on whether we want to advance it.
Gregg Kellogg: If a final report exists it doesn't mean there will be no more. But it creates a published snapshot that can be used as input for the WG.
Gregg Kellogg: Let's triage issues, address them, and move forward. Would be good to have something by June.
Gregg Kellogg: YAML-LD-star won't get to the final report now but we should be tracking JSON-LD-star changes which YAML-LD* is related to. Let's not use YAML specific features at this point. Maybe later.

Topic: Next call

Gregg Kellogg: Next meeting: 10th of May
Gregg Kellogg: Next call tentatively scheduled for May 10th.
Pierre-Antoine Champin: FWIW, May 10 overlaps the AC meeting.
Next meeting on May 3, agenda to be discussed.
Ted Thibodeau Jr.: Difference is considerations of patent issues.
... It's important to do those things to make sure we don't encumber liabilities.