W3C

– DRAFT –
WoT-WG - TD-TF

06 July 2022

Attendees

Present
Cristiano_Aguzzi, Daniel_Peintner, Ege_Korkan, Kaz_Ashimura, Michael_Koster, Michael_McCool, Sebastian_Kaebisch, Tomoaki_Mizushima
Regrets
-
Chair
Sebastian
Scribe
dape

Meeting minutes

Review draft minutes

-> Jun 29

<Sebastian walking over draft minutes>

Sebastian: Minutes look good. Any objections/concerns?
… none -> minutes approved

TD

PR # 1560 Update assertions.csv for testing

https://github.com/w3c/wot-thing-description/pull/1560

Ege: Automatic one..
… IPR issue, but we can ignore... I think

Sebastian: Looks good
… merging

PR # 1561 More in values for apikey security scheme

https://github.com/w3c/wot-thing-description/pull/1561/

Ege: adding cookie, auto, and uri in both JSON schemas

McCool: Looks good

Sebastian: Merging

PR # 1562 remove metadata from example ids

https://github.com/w3c/wot-thing-description/pull/1562

Ege: removed metadata.. in ids
… used online tool to generate v4 ids

McCool: Looks good

Sebastian: Merging

PR # 1549 Add husky pre-commit for rendering

https://github.com/w3c/wot-thing-description/pull/1549

Ege: Looks good

<kaz> Issue 883 - Automate running TD rendering script

Cristiano: does not fix #883
… it does local rendering
… should prevent some errors

Sebastian: Try to resolve conflict
… done

Sebastian: Merging..

Cristiano: reminder.. everyone has to do "npm install" at the root folder of the repository.
… it is mentioned in readme

Issues

Issue # 875 Should it be possible to indicate whether writing a property returns set value?

https://github.com/w3c/wot-thing-description/issues/875

Sebastian: Older topic
… what is getting back from server when writing...
… some implementations echo value
… some don't respond

Ege: It is in the spec now...
… using additionalResponses
… not elegant.. but should work
… maybe we should it fix it "properly" in next version

Cristiano: Agree with Ege...
… wondering about response vs additionalResponses

<kaz> -> 5.3.4.2.1 Mapping op Values to Data Schemas

Sebastian: not sure about additionalResponses and when to use it
… using it for errors?

<kaz> WoT Profile - 6.2.3.1.2 writeproperty

Daniel: not sure what profile does.. expect return value

Sebastian: Changed it.. no value anymore

Ege: properly fixing it needs a way to group response & requests for each type of affordance

Kaz: We should think about potential impact of 1.1 changes to the WoT implementers
… when it comes to existing implementations
… if it is difficult to fix it at the moment, could postpone it to the future version, but should describe the issue as an Editor's note.

Sebastian: Yes, clarify it in next version

Kaz: Consistency check between TD and profile spec might be also useful

Issue # 878 Describing initial connection

https://github.com/w3c/wot-thing-description/issues/878

Sebastian: Some protocols need setup
… discussed a lot...
… but I think it is difficult to solve it in version 1.x
… I suggest to defer to version 2.0

<cris> +1

Issue # 888 Should writing to an observable property trigger new values being pushed

https://github.com/w3c/wot-thing-description/issues/888

Ege: I think we should say there is no such guarantee
… writing should cause new value pushed..
… but this is something a developer needs to do in scripting...
… or might come later

Cristiano: Or the value can be lost...

Ege: Need to do a PR.
… will be non-normative

Issue # 892 HTTP Binding for providing historical events

https://github.com/w3c/wot-thing-description/issues/892

Sebastian: We kind of gave up
… I think Mozilla WebThings has support for it
… but don't think this in profile
… so far we do not have generic approach

Ege: should definitely to that in next charter

Cristiano: +1

Ege: Might also consider storing property value for an entire day

McCool: Historical data should be in next charter

Sebastian: Okay, will label it accordingly

Issue #968 how to find the Form pair for an operation, e.g. the unsubscribe Form for a given subscribe Form

https://github.com/w3c/wot-thing-description/issues/968

Daniel: How to find "correct" unsubscribe

Cristiano: node-wot tries it in best effort approach
… protocol counts most etc

Sebastian: My feeling: depends on how the servient is setup
… might work... but doesn't need to work
… don't think we should make any assumption in TD spec

Ege: Examples in node-wot for examples... have HTTP and CoAP
… could to subscribe on HTTP and unsubribe on CoAP

Cristiano: not sure if there is a good use-case

Ege: Agree
… seems like an Edge case

McCool: Profile defines payload.. unsubscribe should be in

Ege: VERY webhook-like

McCool: I guess we need to leave this up to the implementation
… no simple fix

Kaz: Agree with McCool
… we should invite Zoltan (and maybe also MLagally)

Daniel: Don't think this is a pressing issue

Kaz: Okay, then let's postpone this issue, but we should involve Zoltan and Lagally when we have further discussion.

Sebastian: Okay.. let's defer

<kaz> https://github.com/w3c/wot-thing-description/pull/1517

Static version for TD

Kaz: Copied resources from "4-cr" directory to "4-wd" directory
… we can also move instead of copy

Sebastian: fine by me

Daniel: Didn't get why we changed to WD

Kaz: It is updated WD not CR itself

Sebastian: Want to wait for the CR publication until August. ... need more/better test results

Protocol Bindings

PR #164 Change mjkoster to invited expert

https://github.com/w3c/wot-binding-templates/pull/164

Ege: We agreed on changing M. Kosters affiliation to invited expert
… objections?

Sebastian: Looks good

Kaz: Change itself is fine
… Sometimes some of the Invited Experts would like to mention "former" affiliation
… did you check with Michael Koster?

Koster: No objection
… just saying "Invited Expert" is fine.

Ege: Can add SmartThings as former affiliation in addition to "Invited Expert".
… it is in older documents
… proceed with merging

Other PRs

Ege: Not ready yet, like #156

Testing call

Ege: noticed issue in testing call
… see # 1563

Issue 1563 - Wrong format for tm:required in examples

Ege: it is about the hash sign
… some libraries don't accept #

Sebastian: Strange...

<sebastian> https://tools.ietf.org/id/draft-handrews-json-schema-00.html#rfc.section.9.2

Sebastian: "#" used in JSON schema

Ege: it speaks about uri-encoded JSON pointers

<kaz> RFC 6901 - JavaScript Object Notation (JSON) Pointer

McCool: adds implications whether we need to URI encode it
… we just need local references, no URI

Ege: This would mean we need to fix all the examples
… removing the hash sign

McCool: Correct, seems like a bug in the examples

Kaz: I am OK with fixing the examples
… what was our original intention?

Sebastian: Need to fix examples if they are not compliant

Kaz: no impact on existing implementations?

Ege: Correct
… I don't think this is critical

AOB

Ege: FYI: Next week Sebastian is not available. I will moderate

<kaz> [adjourned]

Minutes manually created (not a transcript), formatted by scribe.perl version 192 (Tue Jun 28 16:55:30 2022 UTC).