22 May 2020



Kaz_Ashimura, Daniel_Peintner, Sebastian_Kaebisch, Ege_Korkan_Taki_Kamiya, Tomoaki_Mizushima


<scribe> Scribe: TK

<scribe> Scribenick: taki

Possible guest from Schneider Electric

Sebastian: A person from Schneider Electric would like to join TD call next time.
... He is considering to join WoT WG.
... Are there any concerns?
... We can make introduction, and he can tell us what he is doing.

Daniel: Is Schneider Electric thinking to join?

Sebastian: They are working on eClass.
... They would like to know what we are discussing.

Kaz: We should talk about this with McCool.
... Inviting once or twice may not be a problem. But regular participation needs invited guest status.

Sebastian: Just one time for now.
... and then they will make evaluation.

Kaz: Sebastian and McCool can decide as co-chairs.

Sebastian: Is it ok for TD TG people? Please consider.
... No objection?

<kaz> (none)

RESOLUTION: TD TF sees no problem inviting Schneider Electric to one of the coming meetings.

PR #896

<inserted> PR 896

Sebastian: I have to manually do it without script.
... Work in progress. Will work next week.

Missing triple resolution in JSON-LD playground for properties forms #905

<inserted> PR 905

Sebastian: I was not sure if this was Playground problem or our definition problem.
... Information about form is missing in N-Quads

<kaz> JSON-LD Playground

Sebastian: This is strange. I suspect this may be JSON-LD side bug.

Daniel: It may be a bug in context file. We need to check with Victor.

<Ege> https://github.com/w3c/wot-thing-description/issues/894

Ege: There was another related issue.

Introduction is missing security component #904

<inserted> Issue 904

Ege: Editorial. In introduction, it says TD has 4 components.
... But Security is missing.
... I think the paragraph was very old. I suggest we add Security.

Sebastian: Rendering script is not working yet.

Managing Dynamically Created Resources in TDs

<kaz> Issue 899

scribenick: kaz

Taki: (goes through the discussion on Issue 899)

Ege: think it could be handled by Protocol Binding
... not really obvious how to deal with ID, though

scribenick: taki

Ege: We need to add a new term in output.

<Ege> https://github.com/w3c/wot-thing-description/issues/302#issuecomment-627500213

Ege is explaining how his proposal works in issue #302...

Ege: Whether we need to use Dynamic TD or not. We first need to decide on it.
... We should first consider static case, then should decide whether we use dynamic TD or not.

Sebastian: We should consider pros and cons between Ege's and Victor's proposal and merge them.
... Multiple consumers initiated actions. In TD, if all actions are visible in TD, can anyone stop other's actions?

Ege: Security mechanism such as token can be used.

Kaz: Mechanism to transfer credential is needed.
... Such additional mechanisms should be considered.

Daniel: If you invoke fade action, how does an implementation know if the action is active or not?
... Consumer needs to check TD update.

Ege: Periodically check action status.

Daniel: How often TD is updated is not known.
... Some action finish shorter others longer.

Ege: Expose Thing script would know that.
... In the case of dynamic TD, it would not know.

Daniel: If both exposed thing scripts and clients die, the added TD part will remain forever.

Ege: When clients do not query anymore, TD is still updated in exposed thing.
... when the action finishes, new forms are removed in dynamic TD case.
... In Mozilla's case, they are kept in memory for some time.
... Programmers can use this option.
... Hypermedia=true option.

Daniel: It may be a bit too complex.

Sebastian: Can Ege create a concrete proposal in HTML document?
... When we have two concrete proposals, we may find the two are very close. We can try in PlugFest.

<kaz> Issue 897

Sebastian: There was a WISHI call.
... Carsten suggested to add constructor class in TDT.
... This is something we can realize by placeholder approach.
... I will write this.

Ege: Use constructor to make complete TD?

Sebastian: Yes.
... TD will be different based on usage. Can use parameters.
... The discussion was at abstract level.

Kaz: The idea is interesting.
... We need to think about TD management mechanism. Object oriented languages such as C++ and Java need to work with some Operation System for actual construction and destruction.

Ege: Scripting API already has produce. Placeholder is not considered yet. It has partial TD.

Sebastian: I suggest to let Carsten to be involved in this dicussion.

Kaz: right. I'm not objecting to this idea itself, but we need to consider those additional points.

HTTP Binding for providing historical events. #892

<kaz> Issue 892

Ege: People display UI using a series of events, for example, Oracle UI.
... Client connect to Thing, and dynamically create UI. Client needs past events to create UI in some cases.

Sebastian: I will create use case issue in Architecture TF.

PlugFest plannings

Sebastian: We have many small features we can test in PlugFests, such as minLength and MaxLength.
... Also CoAP and MQTT vocabularies.
... If we have a good resolution for query action status, we can try that.

Binding Templates

Ege: I plan to work on CoAP vocabulary. A student worked on Modbus vocabulary.
... Is anyone submit PR and gets accepted?

Sebastian: Do you have a concrete PR?

Ege: He has not created a PR.
... Are we allowed accept?

Daniel: We can mark it as not substantial.

Kaz: If it is editorial, it is fine. Protocol binding is not normative.

Ege: It is related to master's thesis.

Kaz: It depends on the content.
... If the content is technical, it may be an issue.

Ege: It will be ttl file.
... A new TTL file.

Kaz: TTL is related to TD, right?

Ege: No. It is about binding templates.

Kaz: Please create the content first.
... Let's review it in next call.

Ege: OK

Sebastian: Klaus will be involved in CoAP vocabulary task.

Ege: I agree we should work with Klaus.

<kaz> [adjourned]

Summary of Action Items

Summary of Resolutions

  1. TD TF sees no problem inviting Schneider Electric to one of the coming meetings.
[End of minutes]

Minutes manually created (not a transcript), formatted by David Booth's scribe.perl version (CVS log)
$Date: 2020/05/25 09:18:17 $