WoT IG Task force on Protocols and APIs -- 02 Jun 2015

<jhund> trackbot, start meeting

<trackbot> Meeting: Web of Things Interest Group Teleconference

<trackbot> Date: 02 June 2015

<jhund> webex: https://mit.webex.com/mit/e.php?MTID=mc3e594ff9016da23da014de4eb63327d

<jhund> do we have a scribe?

<jhund> thank you very much taki!

<scribe> scribe: TK

<scribe> scribeNick: taki

JH: Three agenda points
... Preliminaries
... Architecture and modeling
... and representation of driver model
... How we can fit it into architecture model
... any other topics?

KA: nothing posted yet in IRC

JH: It is now wot-ap channel

<jhund> Agenda: https://lists.w3.org/Archives/Public/public-wot-ig/2015Jun/0002.html

JH: We don't have dates yet for the next F2F. Beginning of July likely.
... Architecture model, and draft primitives of APIs & Protocols. Do we agree we provide them to IG for F2F?
... any other additions or opinions?
... I take we have agreement there.
... Another topic is architecture model.
... Daniel provided first draft, based on Joerg provided on mailing list.
... W3C wiki does not link in images, so we host document there.
... PlantUML code is in the document. Using Graviso.
... I suggested people to review the document and get familiar with the tool.
... any opinions about the tool?

TK: I played around with it a bit, and found it useful for making charts.

JH: The models shown there reflects also what Dave has shown us.
... ... JH introducing what's described in the architecture documents...
... You need github account.
... It can track who edited what.
... History of updates.
... Contributions are welcome.
... Next, Description of driver model

KN: We need to modify two parts. Real implementation of legacy device is one.
... LockServientBT needs additional parts.
... needs to be connected LockServientBT
... I don't think the one described by Siemens as related to driver is not related to driver.

JH: ScriptAPI needs to be there in those two figures, and the APIs will be the same, correct?

KN: Wireless vs. Wired connection is handled by Resources, action, etc box
... From ScriptAPI, they are transparent
... Adapter should show the same API

JH: Whether it is bluetooth or any other connection, the API is the same.
... Do you have github account?

KN: I am not yet familiar with the tool.

<jhund> ACTION: jhund to mail the wiki dump to have easier changes [recorded in http://www.w3.org/2015/06/02-wot-ap-minutes.html#action01]

<trackbot> Created ACTION-8 - Mail the wiki dump to have easier changes [on Johannes Hund - due 2015-06-09].

JH: We will have two kinds of scripting API.
... One on client side and one on servient side.
... two APIs should not change regardless of the protocols to connect devices.
... Resource model, resources and properties. How those model maps to legacy protocols. If you map to HTTP 1.1, events needs polling. We need to collect requirements.

AK: Addressing those devices. Most protocols have it.

JH: We should state what assumption we have for the abstract modeling.
... Not sure if we can say which protocol is suitable over another.

AK: Implicit assumption should be stated explicitly.

JH: "Can this protocol address this in this protocol?" this sort of assumption.
... There is a spot suitable for describe assumptions.

AK: URL-like addressing. Something close enough. Bluetooth qualify as legacy?

JH: Not necessarily. I am not sure, but bluetooth may also have such addressing ability.

AK: That's something we definitely need to look at.

JH: You raised a good point. Mappable or not relates to driver discussion.
... IP-over-legacy, and gateway model. Not mually exclusive.

AK: Ability of addressing and modifying.

JH: Properties, event sources and actions. We assume protocols can represent three kinds of patterns.

AK: There are some protocols that cannot do event sources. i.e. "out of scope"

JH: I suggest guidelines. We address vs. out of scope.
... If you have some legacy protocol. When do you want to have driver?
... When do you want to use WOT protocols?
... Properties, actions and events. Some protocols can map directly, others not.

KN: Device maker needs to fill the gap.

JH: We will make two guidelines.
... Mappability, and whether to use protocol adapter vs driver.
... Other point to discuss here?
... Next meeting will be in one week. 9am central european summer time.
... Thank you for participation, thank you for Nimura-san for architecture model. Thank you Ari for bringing up questions.

- DRAFT -

WoT IG Task force on Protocols and APIs

02 Jun 2015

Attendees

Contents

Summary of Action Items

Scribe.perl diagnostic output