- From: Gunnar Andersson <gandersson@genivi.org>
- Date: Fri, 7 Feb 2020 00:40:57 +0000
- To: Magnus Feuer <mfeuer1@jaguarlandrover.com>, public-automotive@w3.org <public-automotive@w3.org>
Magnus, my friend On Thu, 2020-02-06 at 23:44 +0000, Magnus Feuer wrote: > All, > > We have been exploring an extended VISS protocol that allows for > remote procedure calls to be invoked over the same websocket that > today runs signal pub/sub. No... Just no! I ask you to reconsider this madness, and be reminded (as I have done repeatedly) that there exists a specification that has put considerable time into its development to be complete and correct, and which *also* already has a lot of software written that should not be reinvented from scratch: https://wamp-proto.org/_static/gen/wamp_latest.html > Since we believe this extension may be of use to the wider community, > we would like to explore the possibility of expanding the W3C > standard accordingly. I'm really surprised by this... I was certain that JLR (I mean not only the large car OEM but specifically people close to you in your office/organization) was well aware of the WAMP protocol, since we discussed it quite a lot before during the time VISS was being developed. In some sense it may have been "discovered" by us just a little bit late because VISS specification had already been developed to a large degree, but after that, how can this still be the case? Rather than extending the VISSv1 websocket protocol which is unique to this specification only, for generation 2 I am becoming more and more convinced, and I am ready to advocate for, switching to WAMP entirely! for the WebSocket portion of the specification. (REST part of protocol is still independent). > The proposal, and a working sample implementation, can be found at: > > https://github.com/PDXostc/viss-rpc > > All is open sourced under MPLv2. > > This is in no way a completed spec. Whereas WAMP is for all extents and purposes a completed spec for all the features we are talking about _and_ for additional features (e.g. routed RPC), and for future ideas in their version 2. > Things such as nested arguments > (structs) and callbacks missing, so questions, proposals, and > criticism would be much appreciated. > > If we come to an agreement that this is the right way forward I will > make sure that JLR matures code and documentation as needed to > integrate them into the W3C standard. Again sorry for my bluntness but I think that sounds like a huge waste of time. There are so many open-source WAMP implementations [2] that would be a starting point instead, and that would surely benefit from any available engineering hours for maturing code. Finally, of course, if your investigations have discovered something really new and useful, I'm sure the WAMP development ecosystem would be interested in what you learned! Because the spec is (re)opened for future changes and there has been a *lot* of thinking on that side already about all of this. Reading about their future ideas they are far more advanced in their thinking about advanced features for a messaging middleware of this type. Sincerely, - Gunnar [1] A short(er) overview https://en.wikipedia.org/wiki/Web_Application_Messaging_Protocol [2] Implementations: https://wamp-proto.org/implementations.html [3] Master branch of official spec (in W3C format, even! :) https://wamp-proto.org/_static/gen/wamp_latest.html [4] Overview docs https://wamp-proto.org/spec.html
Received on Friday, 7 February 2020 00:41:01 UTC