W3C

– DRAFT –
(MEETING TITLE)

13 October 2025

Attendees

Present
Andy S, AndyS, Carine, Daniel Beeke, David Habgood, edmond, Holger, HolgerK, SimonW, Thomas (Bergi)
Regrets
Nick
Chair
Eliana
Scribe
DavidHabgood

Meeting minutes

introductions - Daniel Beeke

Introduction - Daniel Beeke

Daniel: looking to contribute to ui work

elianaP: update from Nick - still waiting for horizontal review

Task Forces updates

HolgerK: small core updates, fixes, more time soon to look at tickets.
… last version updated is 1st Sept, expecting more frequent auto publish

elianaP: can look at with caribou after meeting

bergos: PR 555 conformance disallow: was approved, further discussion. Not sure how to proceed, I think OK, looking to merge/create new tickets as appropriate
… two examples have the same IRIs, is this an issue

elianaP: bergos please discuss with Alex and Yousef

AndyS: Alex is currently available due to US gov shutdown

<AndyS> w3c/data-shapes#555

HolgerK: is same IRIs the issue - clashes will happen often, examples are meant to be self contained

bergos: waiting for response from Yousef - will clarify with him directly

elianaP: any updates on profiling? Nick mentioned PR 604, work started on packaging of SHACL; wider working group input will be necessary, relates to presenting SHACL information in general
… profiling taskforce will reach out for input shortly

<AndyS> w3c/data-shapes#604

elianaP: from Nick, call for people to submit other ideas for profiling of SHACL, e.g. profiles based on complexity (216), and any other ideas

elianaP: UI WG updates

edmond: widgets PR open, discussed kinds of inputs to SHACL UI engine, shapes targetting, function that returns set of nodeshapes targeting a focus node. Has been similar discussions elsewhere.
… handling of language literals, browser vs desktop, how should engine determine language, precedence in shapes
… label selection, consider different predicates, sh:name etc.
… using curies, as fallback, local names for IRIs

AndyS: spoke with Livio re. his ideas, he has since written up, it is broader than rules WG

<AndyS> https://github.com/w3c/data-shapes/discussions/603

AndyS: no meeting this week
… re. UI, talking about curies, can be anything, no limitations; prefixed names in turtle/trig etc. are more restrictive
… q re. which parts of the data are targetted by a shape, there must be an interaction between UI and data, any discussion re. protocol? can it work remotely, to e.g. a large database

edmond: targetting is more re. given shapes and data graphs, how does the shape target the focus nodes,

bergos: not covered, suggest more core discussion than UI,
… discussed matching, this can be targets, how to figure out what matches to data

elianaP: scope sounds more general than UI

bergos: might not have people to cover this at present

AndyS: almost a missing document: "SHACL protocols", how to select a set of shapes on the remote server; here is a shapes graph, how do I send it to an engine close to the database

HolgerK: SPARQL has this sort of protocol. Would we be able to start another document - allowed in process?

caribou: depends on whether it fits scope of charter

AndyS: can be a note

HolgerK: need people to work on it

elianaP: too important for note?

AndyS: provides a minimal way of going forward; going from ph1 to ph2, good time to consider. Node expressions mean that, where previously could express everything in SPARQL, this is now harder

caribou: could be an accompanying document, so long as it doesn't contain things that should be normative

AndyS: I will raise issue to start discussion

HolgerK: document will have to span requirements from other documents

AndyS: some are simple; rules is more a symbiotic relationship, need to scope out to see how complex

elianaP: if just note, having overlapping scope is not an issue.

caribou: could create issues with dependencies if other documents are not ready yet
… if just outlining how to use other specs, this is fine for a note, otherwise might need something more

bergos: what should be covered, if implemented with sparql, is this not just an implementation. SHACL SHACL is the only thing we can only really execute in our environment that we've developed. Not sure how many people would rely on SPARQL, rather than an endpoint
… when it spans multiple documents, many feedback loops, things to keep in sync; sceptical at this point on capacity to produce this

AndyS: jena's SHACL engine is the primary way of getting a validation report, send shapes to engine and get validation report

elianaP: Compact Syntax - Jesse and Vladimir not present, will reach out for update

Minutes manually created (not a transcript), formatted by scribe.perl version 246 (Wed Oct 1 15:02:24 2025 UTC).

Diagnostics

Succeeded: i/Daniel:/ Topic: Introduction - Daniel Beeke/

Maybe present: bergos, caribou, Daniel, elianaP

All speakers: AndyS, bergos, caribou, Daniel, edmond, elianaP, HolgerK

Active on IRC: AndyS, bergos, caribou, DavidHabgood, edmond, elianaP, HolgerK, SimonW