W3C

Automotive Working Group Teleconference

09 Jan 2018

Attendees

Present
Urata, Hira, Ulf, Marty, Ted, PatrickL, Gunnar, Hyojin
Regrets
Chair
SV_MEETING_CHAIR
Scribe
ted

Contents


VISS to CR

need to do formal CfC

Ted: Rudi said by email will be sending a call for consensus next week on VISS going to CR
... I can ask for an exception to publish before we recharter and have asked to extend the existing charter for three months

Urata: is it necessary to extend the existing the charter in order to publish?

Ted: it is preferred and will provide an update for that agenda item

Urata: how long, probably 2-3 months?
... do we need to have updated test cases by then?

Ted: we need those and implementation reports to exit CR not enter
... we have a tentative agreement from W3C Management to get an extension for 3 months but it has not been announced to our Advisory Committee yet

Urata: what is the timeline for going to CR?

Ted: after Rudi sends CfC provided we have support from the group and we do not have objection we will publish in January

Urata: there are some remaining issues on the test cases and can work on those in February
... do we have a timeline for getting to REC?

https://www.w3.org/2017/Process-20170301/

Ted: best case scenario is two and a half month

Urata: charter extension should be end of March or April then?

Ted: extension may be long enough to get to REC but we agreed to include completion of VISS in a recharter anyhow
... extension gives us enough time to work on recharter and can publish as well

VIAS

VIAS thread

VIAS Working Draft

VIAS Editor's Draft

https://www.w3.org/2017/11/06-auto-minutes

Ted: this was the first wide review we received for VIAS and it was not favorable

Urata: motivation for starting this thread was at TPAC we discussed VIAS
... in my understanding VIAS is a spec for a JS library which is not a target for W3C standards and should become a WG Note instead
... at F2F I accepted that opinion but not sure everyone is OK so started this thread to confirm and wanted to get others' opinion
... there were a few who responded who preferred not to have it as a Note and want to discuss the reasoning
... it is not very understandable
... it is a JS API but not necessarily a JS Library

https://w3c.github.io/automotive/vehicle_information_api/vehicle_information_api_specification.html#visclient-constructor

in editor's draft VIAS can be on top of VISS, HTTP, MQTT, COAP, IPC

Hira: I sent an email 2 hours ago. VIAS as a WG Note is a surprise to us, because it is quite different from our discussions and conclusion last April in Github #156. I believe most active members already have understood the role of VIAS and its importance from a long time ago. So, “Disposition of Comments on VIAS” paper has been made and posted to Github #253, in order for W3M to understand them.

<urata> https://github.com/w3c/automotive/issues/253

Gunnar: rather than calling it a JS API but rather a programming API in WebIDL
... the question then becomes whether such an abstract API is useful for W3C to publish
... I hear your concern about working on something that may not be viable
... as you said there is not REST/HTTP backend at present

Urata: I added those other protocols as possible future uses. It is a JS API surface
... how it is implemented JS Library or in a browser shouldn't matter nor the backend protocol
... it is about defining API and exposing VSS

Hira: including polyfill

Urata: some people think of VIAS as a wrapper of VISS but that is not necessarily the case. VIAS can be separate and independent

Patrick: having it as a definition of an API is sensible but understand this is different since it is a JS API potentially outside a browser

Gunnar: we could make a separate implementation in other programming languages. it can be defined in WebIDL and implemented in other languages

Patrick: that could count as a second implementation but may not fit with W3C's scope of standards

Ted: reaction was rather strong an example being when HTTP is backend as is an option essentionally GET and POST methods are replaced with get and set

Gunnar and Ted discuss FrancaIDL which is an IDL allowing for implementations (and tooling) for various languages. Since the auto industry is a mix of HTML5 (JS), Qt (C), Android (Java) and other that makes more sense than JS only IDL. W3C may not be favorable towards standardizing it as too broad but will seek clarification. Gunnar offered to take the work to Genivi.

Hira: It seems to me that misunderstanding causes confusion in W3M. I would like Ted to clarify the reason why director cannot support a high-level api spec of VIAS.

Ted: I will send a more detailed explanation of why the TAG and Director are opposed to this specification continuing on recommendation track. [It is not a misunderstanding of W3M but based on their review of the specification itself.]

Explanation sent to member-automotive mailing list.

Summary of Action Items

Summary of Resolutions

[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.152 (CVS log)
$Date: 2018/01/16 17:46:16 $