20 Feb 2019



Kaz_Ashimura, Taki_Kamiya, Manu_Sporny, Matthias_Kovatsch, Kunihiko_Toumura, Michael_McCool, Tetsushi_Matsuda, Tomoaki_Mizushima, Zoltan_Kis, Toru_Kawaguchi, Yosuke_Nakamura, Daniel_Peintner, Ege_Korkan, Ryuichi_Matsukura
Matthias, McCool


<kaz> scribenick: taki

<scribe> Scribe: TK

Matthias: Let's start with quick updates.

IG extension announcement

<kaz> ac announcement (member-only)

Kaz: As I reported last week, WoT IG charter extension was approved and announced. Until the end of June.


Matthias: Workshop dates was changed.

<kaz> updated workshop cfp

Lagally: shifted one week back. To first week of June.
... CfP is getting ready. Program committee is being formed.
... going forward, going to have a weekly call.

<kaz> EasyChair system

Kaz: we had a workshop call. I updated the CfP page, had discussion on how to participate information (we'd like to use EasyChair system to manage position papers).
... will work with the W3C comm team for announcement.

Matthias: Thanks. Let us know when the website is up.

Finalizing the f2f minutes

<kaz> draft minutes from f2f

Matthias: F2F minutes is the next topic.

Kaz: There have been no commemnts.

Matthias: We need resolution.
... any objectton?
... We have a resolution to publish the minutes.

RESOLUTION: the minutes from the Princeton f2f have been approved

Matthias: presentation folder.

<kaz> presentation folder

Matthias: Did everyone put presentations?

<kaz> ACTION: kaz to add links from the minutes to the presentation files

<trackbot> Created ACTION-166 - Add links from the minutes to the presentation files [on Kazuyuki Ashimura - due 2019-02-27].

Lagally: We need links to presentations from the minutes.

Kaz: I will check the sessions, and ping people for presentations.

TestFest status

McCool: TestFest. nobody updated results yet.
... Please knock off remaining ones.

Ege: No updates. I fixed some stuff. no new assertions.

McCool: I removed unique id assetions.
... there are still bunch of at-risk items.
... I will run before TD call.

CR schedule including testing for TD

<inserted> scribenick: kaz

Taki: by the end of Feb for significant ones
... and wide review by TAG, I18N, etc., for 2 weeks from the beginning of March
... and then transition request for CR by the end of March
... spec change last week, e.g., observeThings, subprotocols
... decided to rename interaction pattern to interaction affordance
... victor is looking into schema definition

McCool: 2 more comments
... victor is making change for figures as well
... also regarding TAG review, Sebastian is generating the explainer document and Chairs to review it

Matthias: tx
... sounds like a good plan


<inserted> scribenick: taki

Matthias: Architecture?

Lagally: We just had a call today.
... We went over remaining issue.
... by the end of March, we plan to get the document ready.
... Contribution is needed in the next few weeks.
... Scripting API aspect, please take a look at the document.

McCool: I plan to take a look at the document.

Lagally: Also from security aspect, please review.

How to deal with JSON-LD capability within TD?

Matthias: Sebastian is here now.

Sebastian: There was a question about how we handle TD serialization, and how it relates to JSON-LD.
... What do you think about our current approach?
... we also have JSON-LD transformation section.

<kaz> TD draft - Section 7

Manu: I took a look at those sections and raised three issues.

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

* Issue 442

<manu> https://github.com/w3c/wot-thing-description/issues/442

<manu> JSON-LD Best Practice - Version JSON-LD Context files

Manu: Issue #442
... context can be out of date.
... There is a design pattern.

<manu> publish at URL like this: https://www.w3.org/2019/wot/td

Manu: you publish ontology with pattern like this.

<manu> JSON-LD context file goes here: https://www.w3.org/2019/wot/td/v1

Manu: It gives you a stable vocabulary that apps can rely on without hard-coding.
... Security section section touches on privacy risk, etc. The design pattern address those as well.
... V2 can make changes without breaking deployments.
... It is future-proofing this way.

Sebastian: You mean we should not use generic URL, and we need version information.

Manu: Yes. And we need dates in URL. So it is obvious how old the URL is.

Sebastian: It is good point. we also started to discuss this aspect.
... we also discussed version term. URL solution looks promising.

<manu> Hashlink: https://tools.ietf.org/html/draft-sporny-hashlink-02

Manu: What if Content changes? We are creating a spec to address this using Hash.
... Hash-link specification.

* Issue 443

<manu> Using "http" as a term is strange -- https://github.com/w3c/wot-thing-description/issues/443

Manu: #443 next
... I see "http".

<manu> http:example

<manu> http://example.com/

Manu: "http" led people to think it is a URL.

<manu> {"htv": "http://www.w3.org/2011/http#"}

<manu> {"hto": "http://www.w3.org/2011/http#"}

Manu: Maybe "htv" can avoid such confusion.

Sebastian: Good point.

* Issue 444

<manu> 7.2 Transformation to JSON-LD & RDF seems unnecessary -- https://github.com/w3c/wot-thing-description/issues/444

Manu: #444. about section 7.2.
... You seem to try to achieve round trip transformation. Correct?

Sebastian: It is a desired goal.
... Our TD is not a JSON-LD exactly.
... Because of schedule issue.
... Format based on JSON-LD 1.1. We use @context and @type. We transform to JSON-LD 1.0.
... For semantic processing, we can use transformation.
... as a preprocessing.

McCool: we are concerned with importing into RDF database, for example.
... that is main objective. exporting from database is not main objective.

Matthias: Producing TD can be from data model.
... Ideally TD conforms to JSON-LD.
... JSON-LD 1.1 is currently being worked on to make improvement.

<McCool> Sorry, I will have to drop; plane is boarding; ttyl

Matthias: We can not reference it yet.
... Preprocess is necessary because there are some issues we do not find in JSON-LD 1.1 yet.

<Zakim> manu, you wanted to try to understand what feature is missing to go to valid JSON-LD 1.0

Manu: Curious about those missing issues.
... JSON-LD 1.1 can be described in non-normative section.
... as implementation guide.
... "This section will be updated once 1.1 spec becomes REC"
... W3C allows this kind of language.
... So other group will be on notice.
... in non-normative section.
... Once 1.1 becomes REC, can make it normative.
... What specific feature WoT needs, I need to understand.
... Why WoT can not do with 1.1.
... JSON-LD to JSON transformation. JSON-LD framing is a better way.
... Template using graph to transform to JSON.
... I will ask JSON-LD group to check this.

Matthias: 1.1 looked better for our use.
... @id is re-used with @base hierarchically.

Manu: Tie to 1.1 non-normatively will be the right option.

Sebastian: We should think about it. We did not think about this option.

Sebastion: Normative serialization, we need it.

Manu: There was a precedent. In hind-sight nobody noticed. The spec republished in 4 months with normative statements.
... I suggest to simplify the spec.
... without making extra effort to avoid the issue.

Sebastian: I like the idea, but we need to discuss in WG.

Manu: That's the all feedback I had.
... You will benefit from JSON-LD group's review.
... so they will understand WoT's needs.

Sebastian: Thank you, Manu.

<kaz> latest published WD of TD

Kaz: My position is same as Manu's. JSON-LD 1.1 serialization can be non-normative.
... note that we used to have sections on "JSON serialization" and "JSON-LD 1.1 serialization", e.g., in the latest published WD (shown above)
... we can discuss it again during the TD call on Friday.

Matthias: +1 to Kaz.
... we have raw JSON, and we can preprocess to convert to JSON-LD 1.1.

<manu> +1 to that, sounds good.

Manu's clarification: +1'ing in general to trying to have a section in the specification about "Serialization to JSON-LD", where the focus should be on JSON-LD 1.1 (non-normatively)... where that section becomes normative at some point.

Manu: Pure JSON with @context. This is the only burden for developers. You can hard-code everything else.

Matthias: There is a Media-type issue.

<manu> "application/td+ld+json"

<manu> "application/td+ld+json+jwt"

<manu> "application/td+json"

Manu: The first is a proper IETF way.

<manu> "application/td+ld+json"

Manu: This is a phylosophical debate.

<manu> "application/json-ld"

<manu> "the right way" => "application/td+ld+json"

Manu: That's my suggestion to you.

Matthias: we also want to treat it as a JSON.

TestFest call today?

Matthias: do we have a TestFest call?
... Ege knows?

Ege: There is no big update. We have homework.

Kaz: Manual readme.md is necessary.
... for 03

<kaz> guidelines like: https://github.com/w3c/wot/tree/master/testfest/2019-02-princeton

<kaz> should be added to: https://github.com/w3c/wot/tree/master/testfest/2019-03-online

Ege: OGC people suggested how to document test.

Sebastian: Please share in TD meeting this week.

Ege: OK

Matthias: no TestFest call today.

<kaz> [adjourned]

Summary of Action Items

[NEW] ACTION: kaz to add links from the minutes to the presentation files

Summary of Resolutions

  1. the minutes from the Princeton f2f have been approved
[End of minutes]

Minutes manually created (not a transcript), formatted by David Booth's scribe.perl version 1.154 (CVS log)
$Date: 2019/02/21 06:51:22 $