WoT Scripting API TF Meeting

<scribe> scribe: Yingying

<scribe> scribenick: yingying

[some discussion on the topics today]

[Johannes would like to just go through the github issues as there are limited participants today. Dave suggested some topics like roadmap.]

<jhund> https://github.com/w3c/wot/issues?q=is%3Aopen+is%3Aissue+label%3AScripting

github issues

Johannes: I would like to keep the discussion on github.

-> https://github.com/w3c/wot/issues/247 issue 247

Johannes: in yesterday's OCF call, a company called CEW has very similar model as us. would like to invite them to the IG.

Dave: Just make some guideline and leave designer to make the choice.
... different if you use strong type language or weak type language like javascript.

-> https://github.com/thingweb/wot-typescript-definitons/blob/master/src/wot.d.ts

[Dave talked about the differences for strongly typed languages and dynamic languages like javascript.]

Dave: depends on what model people want for different languages.

Johannes: this is what we think the Promise look like. We need to find the reasonable way to go instead of go into too much details into specific languages.

Dave: I would like to hear what member companies are looking for.

Johannes: go from best practice for javascript and then also can go for C interfaces.

Dave: Question is how do we want to go for IG and what we want to go for WG.

Johannes: should be simple and extensible.
... got lot of comments from Japanese companies during PlugFest.

Dave: I am hoping to get comments from Google and Mozilla from AC review.

Johannes: Ben from Mozilla is still wondering whether we should standardize scripting API.
... it would be good to get them to join the group. if they don't join in the efforts but might be join later on.

Dave: I talked with Dom focusing on the role of javascript and standardization of scripting API. There are some groups in W3C doing on scripting APIs. Device and Sensor API, bluetooth API in community group.

Johannes: main argument is whether we need the server side API. We don't have standardized way to it.
... w3c is in the position to standardize it.
... try to start open source projects with several companies.

Dave: workshop just in starting phase. will keep you in the loop.

Johannes: Masato, I remember you commented on scripting API. do we need a new issue or just use the old ones for comments?

<jhund> Q1.

<jhund> APIs registering / unregistering an event handler to ConsumedThing are different between section 3.3.3 of CP (addListener/removeListner) and Proposal (on/off). Which one do you think is better to choose?

<jhund> Q2.

<jhund> Though "getDescription" and "name" in ConsumedThing are difined in section 3.3.3 of CP, there is no explanation in Proposal about them. Do you have any idea about them?

<jhund> Q3.

<jhund> The parameters of discover() are different between section 3.3.2 of CP (“discover(filter ThingFilter)”) and Proposal (“discover(type string, filter object)”). Which one do you think is better to choose (I’m assuming latter one)?

<jhund> Q4.

<jhund> Do you have any idea about the notation of “discover()” parameters to specify any search query ? An example written in “proposal” shows only repository URL, so in this case I assume that all the TDs in the repository are returned. Also, please let me know if you have any idea to extract such search query to HTTP method (similar to what thingweb-repository implements?)

<jhund> Q5.

<jhund> According to the Proposal, return of ConsumedThing.invokeAction is "a promise that resolves to void, but rejects with an error if occurred", and it seem not to wait the completion of Action. Do you have any idea how to know the progress and the completion of the action? Also, do you have any idea how to cancel the action which has been already invoked?

<jhund> Q6.

<jhund> Which http METHOD do you assume to use when registering/unregistering an event handler? Is POST and DELETE appropriate?

Johannes: let's see if they are already covered by current issues.
... issue 235 covers the first question.

<masato> masato yes

Johannes: for the question 2, name is one field of TD. if you can get the TD you can get it. So question 2 is covered by issue 234.
... however we need better documentation.

<inserted> for question 3,

Johannes: for discovery API, the type is an object. It will return a Promise. Should we go straight forward in this way? or we need to make it more general like type discovery.

-> https://github.com/w3c/wot/issues/240 issue 240

<masato> I cannot hear you

<yingying_> [Johannes went through the issue 240 to explain the discovery.]

<yingying_> Johannes: we also discussed in last meeting about it in issue 235.

<yingying_> issue 235

<yingying_> scribenick: yingying_

[Johannes created a new issue for question 3, issue 267]

-> https://github.com/w3c/wot/issues/267 issue 267

<jhund> seems audio is bad?

Johannes: I also replied your email about the comments. If they are not covered by current issues, we need to create new one.
... would like to discuss them on github because it's easy to track.
... I haven't had my answer yet. Then we should file an issue.
... generally let's discuss them in github.
... suggest we adjourn for today. and will reply you an email and see if they are all covered by github issues.
... thanks everybody for joining.

[adjourned]