The W3C JSON-LD Community Group

Go Back

W3C Logo

JSON-LD WG

Minutes for 2024-08-07

Present
Gregg Kellogg, Benjamin Young, Ted Thibodeau Jr., Niklas Lindström, David I. Lehn
Chair(s)
Benjamin Young
Scribe(s)
Gregg Kellogg, Benjamin Young
Agenda
https://www.w3.org/events/meetings/08563546-30fd-497a-ac16-77f96e96e8c0/20240807T120000/
Topics
  1. Announcements and Introductions
  2. TPAC planning.
  3. YAML-LD
  4. CBOR-LD
  5. JSON-LD Issue Discussion
  6. https://github.com/w3c/json-ld-api/pull/608
Action Items
  1. dlehn to create issue for discussing spec changes related to https://github.com/w3c/json-ld-api/pull/559
Gregg Kellogg is scribing.

Topic: Announcements and Introductions

Wesley Smith: I'm an engineer at Digital Bazaar.

Topic: TPAC planning.

Benjamin Young: I hope everyone can come or dial in.
Benjamin Young: Group meetings are Monday, Tuesday, and Thursday.
... Anyone planning to present anything during breakout?
... wes-smith and I are planing to present on CBOR-LD
Anatoly Scherbakov: How about presenting YAML-LD?
... I won't be in person, but could present remotely.
Benjamin Young: I'd go through the submission process, which should reveal if the technology is in place for remote presentation.
... Deadline is Sept 16th.
Benjamin Young: Spatial Data on the Web WG, JSON-LD WG, JSON for Linking Data CG Joint Meeting, Room TBD, 14:00​–​16:00
... We have some joint meetings, and dedicated group time.
Benjamin Young: Monday: Spatial Data on the Web WG, JSON-LD WG, JSON for Linking Data CG Joint Meeting, Room TBD, 14:00​–​16:00
Benjamin Young: Tuesday: JSON-LD WG, RDF-star WG, JSON for Linking Data CG Joint Meeting, Room TBD, 11:00​–​12:30
... Times are likely/PDT.
Benjamin Young: Thursday...
Benjamin Young: JSON-LD WG, JSON for Linking Data CG Joint Meeting, Room TBD, 09:00​–​10:30
Benjamin Young: JSON-LD WG, JSON for Linking Data CG Joint Meeting, Room TBD, 14:00​–​16:00
Benjamin Young: JSON-LD WG, Web of Things WG Joint Meeting, Room TBD, 16:30​–​18:00
Benjamin Young: Nothing on Friday
... Most of our time will be on Charter, YAML-LD, and joint meetings.
Benjamin Young: Times are in America/Los_Angeles
Benjamin Young: Without objection, we'll eliminate the Thursday 9:00 PDT session.

Topic: YAML-LD

Anatoly Scherbakov: Nothing new on YAML-LD; I've been focused on JSON-LD API issues.
Benjamin Young: Next big thing is to get YAML-LD in the charter as a REC-track item.
Ted Thibodeau Jr.: It might be worth investigating CG rename, from `JSON for Linking Data CG` to `JSON-LD CG` or `JSON for Linked Data CG`
Benjamin Young: We need pierre-antoine to figure out how to rename.

Topic: CBOR-LD

Benjamin Young: Wes-smith recently did a presentation to Verifiable Barcodes talk to the VCWG, which included a lot of CBOR-LD.
Wesley Smith: It's
... "Verifiable Credentials Barcodes".
... The work is about adding crypto security to documents by augmenting existing barcodes which exist on documents already.
... The VCWG initiative augments these.
... This includes drivers license and other official documents.
... CBOR-LD is important because of space constraints.
... This allows the compression we need to fit size constraints.
[Slide presentation ensues]
... The use case is to prevent credential fraud.
... It's small. The crypto standard allows data in the doc and what is referenced to be signed.
... Compressing with CBOR-LD creates a substantially smaller document than the equivalent JSON-LD.
... This constructs an "invertable map" from JSON-LD terms to CBOR-LD terms.
... URLs can be compressed by adding them into the context itself.
... Issuer-specific contexts allow this to be expressed minimally.
... Contexts are processed in a specific way so that it can be verified.
... Verifier reconstructs without using the original map.
... Example 23 shows a map created by processing the contexts into integers.
... Much of the core JSON-LD terms are hardcoded into CBOR-LD (e.g., '@context').
... A Type Table is used to namespace tables to get a range of integers to which it compresses.
... Based on identifying a specific type, I can use that custom mapping for that type.
... You process context documents and compress based on that.
... Specific context IRIs map to well-known integers.
... The mapping must be reconstructable when verifying.
... But, this creates an issue to invert those numbers to find the associated context documents.
... CBOR has tags, which allow semantic information to be associated with CBOR objects.
... We're reserving a tag range for CBOR-LD tags.
... When I see such a value, I can look it up from a spec.
... We use CBOR tag value to find the value for a tag from a registry.
Benjamin Young: Are CBOR-LD processors expected to keep a copy of the registry?
Wesley Smith: Yes, they idea is to keep these as stable as possible.
... The key point is to invert the context string expression.
... I have to invert part of the encoding to find more things to be deoded.
Benjamin Young: The mapping is not looked up, but constructed based on the data. It only includes common mappings, and those that are needed for the specific document.
Wesley Smith: Not everything in the contexts are added to the map.
Gregg Kellogg: Can arbitrary contexts be used/
Wesley Smith: Yes, there's a way to do that.
... There could be a registry entry that allows the issuer to create something, but then needs to hand the custom type-table to the verifier.
... If you understand what to expect, then the arguments are obvious. But, in some cases, you'll need custom mappings.
David I. Lehn: Why are large numbers used for the contexts?
Wesley Smith: Therse is some historical reason for using larger numbers.
David I. Lehn: Should there be a mode which is self-contained?
... So that a processor doesn't need to know too much about external tables?
... There should be some inline mode.
Wesley Smith: There is a default mode without a provided type table. In that case, contexts should up as strings.
David I. Lehn: Any change to a context can mess this up, so contexts should be immutable.
... Maybe needs digests or something else to ensure they're consistent.
Wesley Smith: VC Barcodes is an entry, but there can be other entries for other use cases.

Topic: JSON-LD Issue Discussion

Topic: https://github.com/w3c/json-ld-api/pull/608

https://github.com/w3c/json-ld-api/pull/608 -> Pull Request 608 Resolve the questions about "JSON Serialization" term (by anatoly-scherbakov) [spec:substantive] [ErratumRaised]
PR w3c/json-ld-api#608

Subtopic: https://github.com/w3c/json-ld-api/pull/607

https://github.com/w3c/json-ld-api/pull/607 -> Pull Request 607 Fix test output for html/f004 and add html/f005 (by gkellogg) [test:missing-coverage]
Benjamin Young is scribing.
Gregg Kellogg: My points were to make it clear what to serialize to/from.

Subtopic: https://github.com/w3c/json-ld-api/pull/607

Gregg Kellogg: This was output based on https://github.com/w3c/json-ld-api/pull/606 which has been merged
https://github.com/w3c/json-ld-api/pull/606 -> MERGED Pull Request 606 The default for extractAllScripts in toRdf() should be true. (by gkellogg) [spec:enhancement]
... this cleans up some things following that merge
... dlehn any thoughts?
David I. Lehn: Yeah. this isn't yet handled in jsonld.js
ACTION: dlehn to create issue for discussing spec changes related to https://github.com/w3c/json-ld-api/pull/559
Benjamin Young: We've got about three more calls before TPAC.
... Not scheduled next week, so the next meeting would be Aug 23rd.
... Meeting for Aug 21 moved to Aug 14.
Sorry, I don't know what repository to use.