Meeting minutes
Preliminaries
Draft minutes approval
Sebastian: tackled many PRs
… security, validation, icon links
… profile term
… new terms for dataSchema
… updates in terminology section
… introduce ThingModel namespace
… canonicalization improvements
Sebastian: Call for working draft review
Sebastian: any objections?
<all>: none -> minutes approved
next TD meeting 1h later
Sebastian: Kaz has conflict
… start 1 hour later and have a 1 hour call
Publication plans
Issue 1127
<kaz> Issue 1127 - Publish updated WD for Thing Description 1.1
Sebastian: Started call for review
… collect issues in https://
Sebastian: got already feedback
… e.g., KotisK about trust and security
… should check with McCool
… got comments about broken links
… we do not use final namespaces yet
… for previous drafts we had an intermediary page
Sebastian: can add note making clear that namespaces are not final
Sebastian: any opinion?
Cristiano: Note seems a good idea to me
… temporary URLs sound good too
Sebastian: in previous drafts we used "working" namespaces
… changed to final URL in CR phase
Kaz: 2 comments
… 1. KotisK is official invited expert, we can invite him
… 2. w.r.t. namespace, if we really need it, we can allocate an additional namespace for TM, but the bigger question is about whether we really want to include the TM feature as a normative feature within the CR version of the TD spec.
Sebastian: have to leave for 10 minutes....
<kaz> [ Kaz proposes we talk about the other topics during Sebastian's absence. ]
Propose closing issues
Issue 841
Issue 841 - Add "multipleOf" term to NumberSchema and IntegerSchema
Ege: Any objections to close?
-> none -> proceed with closing
Issue 897
Issue 897 - TDT shall allow to define placeholders
Ege: placeholders are available
… propose to close
-> no objections -> proceed with closing
Issue 1068
Issue 1068 - uriVariables needs to be more limited
Ege: discussion with Cristiano in issue
Cristiano: somewhat related
Ege: Parts are in the spec
… think should be note instead of normal text
Daniel: +1
Cristiano: Suggest to close original issue and create another issue
Ege: makes sense
… created issue 1138
<kaz> Issue 1138 - Making the uriVariables recommendation paragraph an editor's note/
-> no objections to close original issue -> proceed with closing
https://
Kaz: w.r.t. previous issue 1138, note would mean not normative anymore
… if we don't need this sentence we can use ed. note
Ege: I think it is a note for *new* systems..
… kind of a recommendation
Kaz: if the text is included in this section (normative sections) ... and it does not include MUST, SHOULD etc
… not sure if we need to convert it to note
Ege: It is about visibility ... since it is a recommendation
Kaz: Nicer to have a specific section how to use uriVariables
… subsubsection etc
Ege: Seems unnecessary to me
Kaz: contentType might need a dedicated section also
Ege: Makes sense
Cristiano: +1
Ege: makes linking easier also
… will create issue
Issue 957
<kaz> Issue 957 - Unsubscribing and unobserving are not in default values
Ege: Closing issue 957
Issues
Which is better to actuate devices, invoking ACTION or writing PROPERTY? #1020
Issue 1020 - Which is better to actuate devices, invoking ACTION or writing PROPERTY?
Sebastian: MatsukuraSan asks when to use Action and/or Property
Matsukura: original idea is very simple
… table shows standards
… all standards define property
… some define actions
… property defined to actuate physical things
… Echonet Lite just defines properties
… natural to map Echonet Lite properties to WoT properties
… TD is not crystal clear
… suggest to improve the situation
Sebastian: Make clear when to use actions?
… most use-cases can use properties
Cristiano: Do we have a concrete ECHONET example that maps to action?
Kaz: During the vF2F meeting in March, Matsuda-san explained the latest status of ECHONET Lite Web API specification including action/history handling
… page 16 in PDF
… we should wait for their concrete use case description and expected TD binding, etc.
<kaz> Matsuda-san's slides on ECHONET Lit Web API
Cristiano: there are plans to add actions?
Kaz: I believe action is already in work based on Matsuda-san's presentation above, and we were planning to think about possible binding based on some concrete use case and descriptions.
… So I'm a bit confused, and would suggest we should ask ECHONET as a whole about their latest status and their expected direction again.
Matsukura: action in ECHONET is very special
… property is used almost everwhere
… action in ECHONET is a very special case
Sebastian: Would be good to understand why the mapping is not working smoothly
… ECHONET participants in PlugFest would be very good
… suggest to continue discussion in issue
… Matsukura-san, may I ask you to provide a statement about difference between property and action
Matsukura: Sure
Cristiano: similar discussion. Not clear whether write operation returns value?
… not clear in TD spec
… we should try to be more clear
<cris> good question daniel :)
Daniel: writing property in ECHONET, does it return a value
Matsukura: I think just success or failure
… will check though
Sebastian: Discussion in profile about whether PUT returns a value also
… page 20 in PDF shows that payload echoes the value
Kaz: w.r.t. action, see page 21
… suggest to check the latest status with ECHONET
Sebastian: will mention ECHONET procedure in wot-profile PR#77
Kaz: we need a use case and actual data description, also. right?
… given Matsuda-san as the official liaison contact gave the description on the latest status of the ECHONET Lite Web API based on his slides during the vF2F in March and clarified he would continue to be the liaison contact, probably we should ask Matsukura-san to work with Matsuda-san to provide further clarification about ECHONET's latest status and their preferred direction.
<mjk> The OCF API CoAP binding for update also returns the value
Sebastian: Matsukura-san, can you talk with the ECHONET people?
Matsukura: will do
Koster: OCF API CoAP binding for update also returns the value
… use POST (instead of PUT) .. because of partial updates
Cristiano: is this a profile issue or is this a broader issue?
… seems more an architecture issue
… we seem to have use-cases
Koster: feature of TD whether to expect a return value
Ege: different aspect: action does not need input/output
… write property is different
Cristiano: Agree, different design choice
Koster: for the industry we need some flexibility
… model *looks* like property but is action...
… different return value might mean it is an action
Cristiano: +1
Ege: +1
<ryuichi> +1
Sebastian: clarifications with ECHONET are useful also
Koster: it is in some SDF discussions also ...
Sebastian: actions to me was always somethings that takes time
Koster: we need to figure out best practices
Sebastian: Okay... let's get back to this the next week
… Thanks, Matsukura-san!
Ege: FYI: related issue: observe vs event
Sebastian: Agree, same kind of category
Open PRs
More fixes to canonicalization #1129
https://
McCool: new rules
… talked with Victor also
… security objects were inlined
… re-frame from RDF reversal needs to recreate name
… base was an issue also
… base cannot be changed in Canonicalization process
… cleaning up corner cases
Sebastian: Should we merge it?
McCool: not controversial .. but no review yet
… created wd-update-candidate branch
… review should be done in this branch
Sebastian: rendered version in main branch only
Daniel: not sure if this is overly comlpicated.. could hold back merging
McCool: should not change reviewed document
… typos could be merged
… bigger PRs should not be merged
fix: ReSpec rendering issue #1132
https://
Daniel: just escapes the "{{"
Sebastian: Suggest to merge
… no objections -> merged
fix: some typos and misspellings #1133
https://
Sebastian: no concerns --> merging
fix: tweak minor example bugs #1136
https://
Daniel: just fixes examples
Sebastian: no objections -> merging
Comments - Call for Review of WoT Thing Description 1.1 specification and resolution to publish update #1137
https://
Daniel: bug w.r.t. MultiLanguage
Sebastian: Yes, we should fix that
McCool: suggest hold of creating a branch
… merge my PRs later
Sebastian: makes sense
McCool: Will work on my PRs in the meantime
Sebastian: I will create a PR for issue 1137
… I also suggest to update abstract, see https://
McCool: Seems more than just a bugfix
… maybe discuss it in main call
Sebastian: plan to use initial example for ThingModel
McCool: main call discussion next week?
Sebastian: Agree
Issue 1127 (revisited)
Kaz: We need to go back to issue 1127
… missed some points
… need to clarify what is normative
… e.g., ThingModel
McCool: I think ThingModel should be normative
Kaz: either way is fine. We just need to be clear
McCool: mark it as "at risk" is an alternative
Sebastian: FYI: Editdor project is implementing ThingModel already
McCool: I am also not concerned
Sebastian: It used to be part of the Appendiex, and I'm not sure if it should be included in the TD spec, though
kaz: in that case, we can identify the Thing Model feature as a feature at-risk anyway
… if we can get sufficient implementations, that's fine and great
… I wanted to make sure about this point before our sending out the call for review request
Sebastian: ok
[adjourned]