W3C

Automotive Working Group Teleconference

07 Jul 2020

Attendees

Present
Glenn, Harjot, Ulf, Adnan, Benjamin, Gunnar, Christian, Daniel
Regrets
MagnusF, Peter
Chair
SV_MEETING_CHAIR
Scribe
ted

Contents


zak

<scribe> scribenick: ted

VSS instances

https://github.com/GENIVI/vehicle_signal_specification/pull/164

Daniel: encourage people to read the proposal in the pull request and what Sebastian has contributed
... we can discuss more next week

Gunnar: I could clear up my view on JSON from previous call
... we should have a single definition of what VSS is, content and semantics
... no objection for using JSON. today we have a near-exact mapping currently from YAML to JSON
... when we discuss the csv format, the main concern is the full path signal name
... it is not supported today
... Gen2 is based on CSV file with plain path as to how you address things

Ulf: I think so too

Gunnar: concerned about complete reliance on JSON

Daniel: not decided how to do it in tooling. question is how it looks in the end and how to go forward
... today we have VISS and start of Gen2. we also have graphql, VSSo and number of open questions around Android

Gunnar: core question is do we have a new concept in the source (YAML) model and does it survive translation to other formats including JSON
... to progress that we should first clarify how to use that knowledge
... I am always concerned about creating a specialized solution for something and prefer generalized
... not sure you can do so much with instances, not clear it is a position and cannot programmatically know it is a position

Daniel: it all belongs to the same concept. my desire was to have position at the very end and flatten it out
... would like to be able to create a dictionary out of the instances

Gunnar: I am partial to having it as a real part of the model and would prefer something cleaner
... when do you need to know instance is same concept (and there are others)

Daniel: we need some more reasoning and examples behind it

Gunnar: proposal now has instances
... maybe each protocol (format) decides its preferred representation
... I keep falling back to YAML being source of truth
... either need to adopt instance concept in Gen2 or translate for path

Daniel: csv should give idea on how to address instances directly

Ted: for VISS, our examples were based on VSS1, for Gen2 we need to be able to explain how to understand positioning, find things and represent with HTTP path and possibly querystring parameters or other to be able to address a specific instance or get back all instances of eg door

Gunnar: you can do that with wildcards. not sure query syntax creates any additional availability

Daniel: you want a REST interface to be able to get to individual doors

Gunnar: if I just want to get concept of door out of model...

Daniel: you can get everything from an instance

Gunnar: if I put a * or something else in path, I want concept of door instead of a specific instance

Ted: presently we enjoy a very clear and direct translation between YAML and JSON. appreciate positioning might be handled better for Graph and other uses and in my opinion fine if there are deviations among the formats although would prefer consistency

Gunnar: JSON translation could be done still but maybe GENIVI VSS only produces YAML and different formats are created for their various uses
... you get YAML from VSS and can do whatever you want with it. we cannot anticipate others' desired translations
... we are doing that presently for CSV and JSON but it could transition
... you could load YAML just as well in Go code

Ted: there will certainly be people that come up with their own representations for different formats/protocols outside of the VSS repo. to the extent we can provide authoritative formats for the ones we care about the better.
... one idea would to also have the VSS repo act as a registry for alternate formats maintained elsewhere

TPAC

W3C TPAC schedule

W3C Groups

Ted: [summary of how TPACs work with joint group meetings and unconference style breakout sessions]
... it will be virtual this year for obvious reasons, want your input on who else besides webrtc, wot, sdwig that may be of potential interest

Gunnar: wonder about input on security aspects

Ted: yes and can try to bring over some hopefully for that

[adjourned]

Summary of Action Items

Summary of Resolutions

[End of minutes]

Minutes manually created (not a transcript), formatted by David Booth's scribe.perl version (CVS log)
$Date: 2020/07/07 19:05:16 $