Re: RPC protocols and the path forward.

Hi T,

Apologies, in giving history of the group I didn't say we are
essentially done with VISS. As such I wouldn't necessarily worry about
trying to align WAMP with VISS, just Gen2. Also we decided yesterday
WAMP is a strong contender for RPC proposal we want to have
architecturally similar to Gen2.

VISS has shipped and in some production vehicles. It is fairly
lightweight and we do not have plans to evolve it, the as yet named
'Gen2' is its successor. A name in consideration is VISS version 2,
mentioning to try to avoid future confusion.

Another somewhat confusing part is similar names between VSS and VISS.

Vehicle Signals Specification, VSS is the underlying data model that
VISS API exposes. It is over at GENIVI and we work in close
collaboration with them.

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

Vehicle Information Service Specification, VISS as mentioned is the
API.

https://www.w3.org/TR/vehicle-information-service/

So the naming you do, inverting domain names as done for Java clases
and elsewhere, for the data model would probably be more appropriate
starting with com.genivi. While W3C has the domain w3c.org it largely
uses w3.org, our Java classes for whatever reason use org.w3c.  I know
it is just an initial assessment but worth clearing up.

Ah, fun with names.

Regards,

T
 
On Mon, 2020-09-28 at 23:12 +0200, Tobias Oberstein wrote:
> > Tobias offered to give this group an overview of WAMP and to take a
> > look at our Gen2 effort to give his thoughts on how they may
> > possibly
> > merge.
> > 
> > https://w3c.github.io/automotive/spec/Gen2_Core.html
> > https://w3c.github.io/automotive/spec/Gen2_Transport.html
> 
> fwiw, I've looked through these 2 specs (and the VSS). anticipating
> the 
> actual wamp presentation, here are my first thoughts on mapping
> vss/gen2 
> to wamp specifically:
> 
> https://gist.github.com/oberstet/efaba6eaaf4170b39550b25d6a28a5cd
> 
> but an outright mapping as I understood is not necessarily a goal. 
> anyways, might be useful nevertheless ..
> 
> ---
> 
> a naive question though:
> 
> did anyone actually ever implement a working prototype 
> VSS+gen2core+gen2transport?
> 
> did 2 independent prototypes ever talk to each other?
> 
> did anyone put on an app developer hat and scrutinized whether the
> API 
> and all that "feels right" (for an app developer)?
> 
> ^ as in: "rough consensus PLUS working code" ;)
> 
> IMO spec authors should both be the first implementors and the first 
> users. to make sure, the stuff actually works and fits needs.
> 
-- 
Ted Guild <ted@w3.org>
W3C Automotive Lead
https://www.w3.org/auto