Re: Proposal: Web Thing Protocol Community Group

Hi Ege,

On Tue, 17 Sep 2019 at 07:45, Ege Korkan <ege.korkan@tum.de> wrote:

> I am still having the opinion that this topic could be covered in the
> protocol bindings task force, which would be more than happy to have you
> onboard!
>
I agree that it could, but so far it hasn't. This Community Group is
intended to accelerate that work through a focused effort which invites
broader participation from contributors who for whatever reason (e.g. the
high barrier to entry or disagreeing with the charter) have chosen not to
join the Working Group. It is proposed in the spirit of co-operation with
the existing Working Group & Interest Group.

> There we can talk of defining the default http/ws protocol binding which
> could be named as the Web Thing Protocol. We can have a separate document
> but this would look like "yet another standard" from my point of view.
>
That is the intention. I see there being two important normative
specifications for the Web of Things (they could be part of one document,
or two separate documents):

   1. Web Thing Description - a file format for describing the capabilities
   of devices
   2. Web Thing Protocol - a protocol for communicating with devices over
   the web

So far the Working Group has done a great job of defining the former, but
the latter has not yet emerged.

> In the meantime, I have found this draft from you:
> https://iot.mozilla.org/wot/
>
> It seems to be up to date and referencing the W3C WoT Thing Description.
> However, it is still using the Mozilla way of doing things and not really
> using the W3C WoT Thing Description and defining a new "Thing Description".
> Moreover, I couldn't find a way to write issues or discuss things, thus the
> email. Is this a preliminary version until the Community Group is formed?
>
That specification has existed since November 2016 (GitHub repo including
issue tracking here <https://github.com/mozilla-iot/wot>) and, as the
"Status of This Document" section explains, "continues to be maintained to
reflect the current API implemented by Mozilla's own open source Web of
Things implementation".

The "Web Thing Description" section of the specification has gradually been
converging with the Working Group's "WoT Thing Description" specification
over time and is intended to eventually be superseded by a W3C
recommendation once one is agreed. The main significant remaining
difference between the two is that of "forms" (declarative hypermedia
controls) vs. "links" (endpoints for an externally specified protocol).

The "Web Thing REST API" and "Web Thing WebSocket API" sections of the
Mozilla specification cover areas not yet fully covered by a W3C
specification, which is what the new Web Thing Protocol Community Group is
intended to address.

There has recently been some work in the Working Group on defining a the
beginnings of a default HTTP protocol binding
<https://w3c.github.io/wot-thing-description/#http-binding-assertions>
through defaults in the Thing Description specification. There is also
ongoing discussion <https://github.com/w3c/wot/issues/873> regarding a
potential "profile" mechanism and/or "implementation view spec" as work
items in the next Working Group Charter
<https://cdn.statically.io/gh/mmccool/wot/wg-charter-draft/charters/wot-wg-charter-draft-2019.html?env=dev>,
which may or may not provide a more complete default HTTP protocol binding.
If that were the case then the Web Thing Protocol Community Group may
simply reference that work. If not then the Web Thing Protocol Community
Group will work on this itself (as described in its charter).

There is currently no work in the Working Group on defining a WebSocket
sub-protocol for the Web of Things. This is something the Web Thing
Protocol community group will definitely work on.

I hope this helps clarify things.

Regards

Ben