Schemata Follow-up

W3C Breakout Day

March 12, 2024

​

Ege Korkan, Mahda Noura

Logistics

• W3C Calendar Entry
• Please all join IRC at https://irc.w3.org/?channels=%23schemata-discussion
• Then type present+ to check in.
• Type q+ to raise hand
• We will take minutes there manually.
• Scribe will be Jan
• Quickly introduce yourself before speaking (now if there are not too many people)
• These slides are public. Link.

W3C Breakout Day | March 12, 2024 | Schemata | Ege Korkan & Mahda Noura

Participation Policies

• Antitrust and competition policy
• W3C acts in a pro-competitive way that is ensured by this document.
• Positive Work Environment at W3C: Code of Ethics and Professional Conduct
• Promote high standards of professional practice to ensure a positive work environment

W3C Breakout Day | March 12, 2024 | Schemata | Ege Korkan & Mahda Noura

Bringing some Context

Nature of this session:

• Brief presentation
• Discussion

Goal:

• Share experience
• Find a place within W3C for this type of discussion

Required background:

• Basic understanding of JSON-LD, Ontologies, Schema languages

4

W3C Breakout Day | March 12, 2024 | Schemata | Ege Korkan & Mahda Noura

Have we met before?

Were you in the previous session of this?

​

​

​

​

​

​

If not, please write "new" in IRC :)

W3C Breakout Day | March 12, 2024 | Schemata | Ege Korkan & Mahda Noura

In case you haven't been there

Slides from last time:

• Introduction from Pierre-Antoine Champin: https://perso.liris.cnrs.fr/pierre-antoine.champin/2023/TPAC-Schemata/
• Web of Things TD Use Case from Ege Korkan: https://github.com/w3c/wot/blob/main/PRESENTATIONS/2023-09-tpac/2023-09-13-Breakout-Schemata-TD_Single_Source_of_Truth.pdf
• Minutes: https://www.w3.org/2023/09/13-schemata-minutes.html
• Quick Summary

W3C Breakout Day | March 12, 2024 | Schemata | Ege Korkan & Mahda Noura

Quick Summary

W3C Breakout Day | March 12, 2024 | Schemata | Ege Korkan & Mahda Noura

Quick Summary

Web of Things TD Task Force needs to manage the following:

• Spec document, which contains vocabulary terms and information model
• Ontology documents
• SHACL shapes
• JSON Schema
• Type and Class Definitions (for now only TypeScript)
• Test cases
• Examples

All of them need publication procedure.

Soon, each binding will need the same and anyone should be able to do it…

W3C Breakout Day | March 12, 2024 | Schemata | Ege Korkan & Mahda Noura

Two Presentations

We knew that there are existing approaches to solve this and we have invited them for presentations:

• LinkML by Chris Mungall
• TreeLDR by Timothée Haudebourg

W3C Breakout Day | March 12, 2024 | Schemata | Ege Korkan & Mahda Noura

What is in the Follow-up though?

• Results of the analysis of approaches/tools by the WoT WG
• Versioning, packaging and serving resources
• Continuation of the work
1. We should reserve the last 10 mins, please remind me if forgotten

W3C Breakout Day | March 12, 2024 | Schemata | Ege Korkan & Mahda Noura

Analysis by WoT WG (mostly Mahda Noura)

Two-fold procedure:

1. What we were doing so far
2. What can we adopt to make our lives easier

​

All information is contained at https://github.com/w3c/wot-thing-description/tree/main/toolchain for now.

W3C Breakout Day | March 12, 2024 | Schemata | Ege Korkan & Mahda Noura

What we were doing so far

Not proud but it sort of works :)

W3C Breakout Day | March 12, 2024 | Schemata | Ege Korkan & Mahda Noura

What can we adopt to make our lives easier

W3C Breakout Day | March 12, 2024 | Schemata | Ege Korkan & Mahda Noura

Any Feedback is Appreciated via Issues

Anyone can do so at https://github.com/w3c/wot-thing-description/issues

We already got one from Vladimir Alexiev, thank you!

​

For the WoT TD purposes, LinkML seems like a good fit at the moment but we haven’t decided yet.

W3C Breakout Day | March 12, 2024 | Schemata | Ege Korkan & Mahda Noura

Ok, let’s say we picked one, what is next?

You have generated all those resources.

Now how do you version, package, serve all of them?

W3C Breakout Day | March 12, 2024 | Schemata | Ege Korkan & Mahda Noura

Manual Way

Create a GitHub repo, enables Pages but only serve resources:

https://github.com/w3c/wot-resources is used at the moment by WoT WG. We push resources there, team contact adapts redirection settings. No versioning at the moment.

​

It is similar to uploading software to a web server,

like we did before package managers were a common thing,

20 years ago…

​

Too slow for prerelease cycle (before REC) and ambiguous for post REC.

W3C Breakout Day | March 12, 2024 | Schemata | Ege Korkan & Mahda Noura

Can we do it better?

We are discussing versioning rules. Draft PR

How about tooling?

• OS/Distribution-specific Package Managers
• Programming Language Package Managers
• Some early work from Klaus Hartke via using npm and package.json format
• Container Registries
• Custom-built Tooling

​

​

W3C Breakout Day | March 12, 2024 | Schemata | Ege Korkan & Mahda Noura

Any Ideas?

W3C Breakout Day | March 12, 2024 | Schemata | Ege Korkan & Mahda Noura

Where should the discussion continue?

In TPAC23, there was interest to form a CG:

​

W3C Breakout Day | March 12, 2024 | Schemata | Ege Korkan & Mahda Noura

Where should the discussion continue?

Personal opinion: there is not necessarily need for standardization but there is at least need for discussion and exchange, finding best practices.

Other opinions?

W3C Breakout Day | March 12, 2024 | Schemata | Ege Korkan & Mahda Noura

Discussion Points

Tooling:

• Vladimir: Mix and match solutions that are not optimal.
• AB and VA: Not losing data or meaning when translating one representation to another is very important

Versioning, Serving, Packaging:

​

Where to continue:

W3C Breakout Day | March 12, 2024 | Schemata | Ege Korkan & Mahda Noura

Relevant Communities

Trade Transparency group

​

Discussion in YAML-LD: https://github.com/json-ld/yaml-ld/issues/19

​

https://github.com/w3c-ccg/traceability-vocab/issues/295

VA: SHACL to UI (in rdfjs) will be using the most common subset of SHACL.

VA: Existing GraphQL implementations that use RDF -> Link

W3C Breakout Day | March 12, 2024 | Schemata | Ege Korkan & Mahda Noura

Other Comments

Netflix ontology service (Alexandre B):

• uses SHACL. How will authoring work?
• SHACL is way too powerful -> steep learning curve
• How to get meaningful ontologies out of SHACL
• Discoverability is an open topic
• The turtle users don’t know that they write ttl

Ivan:

• Common problem
• SHACL and shex cannot model datasets
• RDF usage is quite different between implementations. RDF 1.2 will intensify that
• JSON-LD has a double nature that creates the root of the issue
• Where is the discrepancy between these formats -> needs to be clarified first

W3C Breakout Day | March 12, 2024 | Schemata | Ege Korkan & Mahda Noura

Check-out (to be extended in the end of the meeting)

A summary before the discussion ends:

• Main points of discussion, consensus, or disagreement?
1. Consensus on the annoyance of the problem and relevance to different communities (not just WoT)
• What are the next steps?
• Creating a more clear picture of the problem. Catalog.
• Dissemination
• (proposal) focus CGs to work out specific issues: a) SHEX-SHACL mapping, b) RDF-GraphQL best practices, c) maybe YAML/YAML-LD syntax for mixing schemata
• Who is responsible for carrying them out? (Could be a person from the session, or a group where work is ongoing, a new community group, the staff, etc.)
• Each person here -> Noting these problems and dissemination.

​

W3C Breakout Day | March 12, 2024 | Schemata | Ege Korkan & Mahda Noura