Automotive Working Group Teleconference

<scribe> Scribe: Ted

<scribe> scribenick: ted

Pull requests (315, 320)

PR315

Daniel: I lessened the wording to should as I agree we may want to handle multiple domains separately
... it is a compromise for now and new issue can be raised if there are still concerns

Ulf: I agree, like the wording and it can now handle multiple domains

Gunnar: does this allow different formats or other variations?

Daniel: I took from the issue and discussion the interest in being able to handle other domains with this same protocol
... there was resistance to having a single tree so changed from shall to should

Ulf: as previously discussed, if most of the PR is good but may have minor issues we should accept the request and raise those issues independently for further discussion

Peter: I agree with Ulf

Ted: I am good with it too, cannot speak for PatrickL but as we said we can raise explicit concerns as new issues and not hold back the PR

[recap of 315 as dis cussed before PatrickL joined above. the following took place later in the call and moved up here]

https://github.com/w3c/automotive/issues/310

https://github.com/w3c/automotive/pull/311

https://github.com/w3c/automotive/commit/7f3ddfb2f8cf09b09b403f65ac6e5787dc03c10a

Daniel: combined with should instead of must is start of a compromise

PatrickL: but why compromise?

Daniel: we need to get concrete with taxonomy being served, what the protocol is designed for it weakens it
... there are plenty of vague protocols out there but we want more structure here
... having domains connected to the vehicle and how you can link to them it would be a stronger point

PatrickL: I will review the discussion in 306

<PatrickLue> " The single entry point and master of the Gen2 Multi Domain Taxonomy SHALL be the <a href="#vehicledomain">Vehicle Domain Taxonomy</a>."

Magnus: I agree with Patrick that we should explain we can use different taxonomies for different domains

<PatrickLue> from line 89/90

Magnus: I agree with Daniel that we should have something to point to
... they should be encouraged to use VSS and agree

PatrickL: SHALL is still in the pull request

[still stuck on single tree]

Ulf: my updated slides help with this
... service manager can add new branches and/or provide links instead

Ted: so duplicate URI space but some may consider an optimization, also distributes potential integration points within vehicle
... again suggest we discuss registry

Daniel: if you don't define relationship to the vehicle why have it there

<PatrickLue> Gen2 is not only for the vehicle domain

Daniel: should think of the data model and beyond VSS but tie

#320

Ulf: the commits should help with the filtering issue and have a corresponding demo if people want to see

Peter: I think we would love to see the demo

[software architecture diagram overview]

https://github.com/MEAE-GOT/W3C_VehicleSignalInterfaceImpl/

[corresponding method requests enumerated]

@@requests

Ulf: methods from websockets are 1:1 corresponding to http
... server on separate host. i'm using a JS client in html page as app
... need to tell it where to find server, it connects and then you can make your query eg isOpen on all four doors

[demoing form and backend requests]

[multiple subscription streams merged, dropping one...]

[wildcards working anywhere other than at the root path]

Gunnar: the subscription interval request didn't feel exactly right to me
... this syntax is same as queries and wonder if VISS didn't have this?

Ulf: this is copied from VISS, it is supported as a key/value pair in the request
... this is different syntax. advantage is you can do the same over websockets and http

Gunnar: we should be conscious of backwards compatibility. are you suggesting we provide the old version as well

Ulf: correct, would be without old method but feel it is worth the trade off

PatrickL: how about putting it in POST method instead of URL?

Ulf: that is a possibility and considered it
... I can agree it is a matter of taste

Ted: generally anything affecting change preferred as POST at W3C

PatrickL: depends on what is considered change, as you are getting results impacted it might be better as POST
... we had similar experience in ViWi

Ulf: we should probably allow for review period

Peter: demo was a great introduction

Holidays

clearly not the meeting on 24th nor 31st

VISS

Ted: JLR may want to make a couple small changes to their report, Volvo should consider doing one too
… closed small bug that Ulf reviewed. other than that there is VSS version number which is an issue in that repo and then we should advance to Proposed REC
... probably should just include VSS version number in root