"You say Schemata, I say Schemas" - TPAC 2023

13 September 2023


AaronCoburn, aciortea, atai, Bert, Cristiano_Aguzzi, dape, David_Ezell, decentralgabe, Dingwei, doerthe, Dominik_T, Ege, ethieblin, glomb, ktk, LaurensDebackere, michielbdejong, Minyong_Li, pchampin, phila
Ege Korkan, Pierre-Antoine Champin
Ege, pchampin

Meeting minutes

<ChrisMungall> What @pierre is saying resonates with me, I love OWL, but I frequently see it mis-used for describing data and not the world

the link for the presentation: https://docs.google.com/presentation/d/1iKg0K_tIFdZJ2DBLiHIqjAc5ydo7ifUYYkkkf2Nc_D4/edit?usp=sharing

<ChrisMungall> TreeLDR looks great, I wasn't aware of it

<Zakim> question, you wanted to comment on governance and release plan/stability

<Dominik_T> TreeLDR reminds me a lot of PG-Schema formalism: https://arxiv.org/pdf/2211.10962.pdf

cris: thank you for the presentations. it sounds great and starred it
… it sounds too good to be true. Can you see limitations for a complicated schema

cm: We can incorporate such additional expressivity in next releases

th: all the concepts of owl are possible but maybe not everything you can do in json schema can be modeled

ktk: thank you, treeldr looks interesting. we wanted to do something similar
… we also generated SPARQL queries and also html forms
… we have a chrome based RDF debugger. If you do validation, you do not need a debugger anymore
… you can create ridiculuosly long sparql queries
… in typescript you compile, not runtime data
… do you have a frontend
… can I create from RDF DSL generation?

th: not at the moment

dp: I am impressed. How is the stability, are you planning to standardize?

th: we are in early stages

<cris_> treeLDR: spruceid/treeldr

<cris_> linkml: linkml/linkml

th: we can think of standardization but we are not there yet

cm: we have monthly community meetings

Ege: recently XXX proposed different vocabularies for validation.

cm: we would be interested to explore how it can be standardized

Ege: Chris, you mentionned that some things were not possible in JSON-Schema. What did you have in mind?

ChrisMungall: things like referential integrity are not possible in JSON-Schema. Only a problem if you only use JSON-Schema.

<Zakim> Ege, you wanted to governance and release plan/stability of treeldr

phila: I did not catch what you said about dynamic enumerations?
… have you solved that?

cm: I do not know if we fixed that

<phila> s/dynamic xxx/dynamic enumeration. Have you solved that?/

ca: I have worked with Prisma recently, have you heard of it?

<Zakim> phila, you wanted to ask about dynamic enumeration

ca: it is common in Node.js runtime workers since it works well with react etc.

<cris_> prisma: https://www.prisma.io/

mm: wot uses schema to validate td but also for describing payload
… what is a good place for linkml, can you use it in embedded devices

cm: I do not know about iot but generally it is used with DBs (mongodb etc.)

mm: we do different levels of correctness checking in wot

cm: some languages also have warnings/recommendation so that can be picked up by linters

cm: you can also have profiles

pc: interested about profiles

pc: what are your thoughts on the motivation I have put in the beginning
… cardinality etc.

th: layouts would be the solution

cm: it is a closed model. I would not recommend for open applications

<Zakim> phila, you wanted to talk about profiles

<McCool_> not sure will have time... but what about RDF-star?

phila: I want to pick up on profiles. Profile negotation is an old thing but it getting picked up again
… asking for json ld but using that shacl file

<phila> https://www.w3.org/2022/06/dx-wg-charter.html

phila: it would be good to check with data exchange wg

pc: topic of standardization has been raised
… we can think of a cg

<glomb> +1

<dezell> +1

<aciortea> +1

<ChrisMungall> +1

<Dominik_T> +1

<cris_> +1

<ktk> 0

<Bert> 0

<pchampin> STRAWPALL: should we create a CG on this topic

<mahda-noura> +1


<McCool_> +1

<Dominik_T> +1

pc: we can make a proposal for a cg based on the description of this proposal
… staff should not make it but we can exchange after the session

ktk: tooling wins in the end with such things

pc: we can use it for idea exchange

Ege_: would that fit into an existing group as well?

pc: we should check
… and maybe meet in a new group next time

Minutes manually created (not a transcript), formatted by scribe.perl version 221 (Fri Jul 21 14:01:30 2023 UTC).


Succeeded: s/pa/phila/

Succeeded: s/XXX/enumerations

Failed: s/dynamic xxx/dynamic enumeration. Have you solved that?/

Succeeded: s/pa:/phila:

Maybe present: ca, ChrisMungall, cm, cris, dp, Ege_, mm, pc, th

All speakers: ca, ChrisMungall, cm, cris, dp, Ege, Ege_, ktk, mm, pc, phila, th

Active on IRC: aciortea, Bert, ChrisMungall, cris_, dape, dezell, doerthe, Dominik_T, Ege, Ege_, glomb, ktk, mahda-noura, McCool_, minyongli, pchampin, phila, tidoust