Strawman Proposal: Web Thing Protocol - WebSocket Sub-protocol

Dear members of the Web Thing Protocol Community Group,

Following the publication of our Web Thing Protocol Use Cases & Requirements
<https://www.w3.org/community/reports/web-thing-protocol/CG-FINAL-web-thing-protocol-requirements-20231101/>
report, I am excited to kick off work on the core deliverables
<https://w3c.github.io/web-thing-protocol/charter/#deliverables> in our
charter.

The first deliverable I suggest we tackle is a WebSocket sub-protocol
specification.

I hope you don't mind me kicking off the discussion with a* strawm
<https://docs.google.com/document/d/1KWv-aQfMgsqBFg0v4rVqzcVvzzisC7y4X4CMUYGc8rE/edit?usp=sharing>*
*a*
<https://docs.google.com/document/d/1KWv-aQfMgsqBFg0v4rVqzcVvzzisC7y4X4CMUYGc8rE/edit?usp=sharing>*an
proposal
<https://docs.google.com/document/d/1KWv-aQfMgsqBFg0v4rVqzcVvzzisC7y4X4CMUYGc8rE/edit?usp=sharing>*.
This is a quick (but quite detailed) draft document in a Google Doc
outlining a possible WebSocket sub-protocol design approach.

This design is informed by implementation experience from the relatively
primitive WebSocket API <https://webthings.io/api/#web-thing-websocket-api>
used in the WebThings platform, but is a more sophisticated design which
includes the full set of operations which have now been defined in the Web
of Things interaction model. It is also flawed in various ways.

In the proposal I have defined a collection of message payload formats for
the full set of WoT operations. For Property related operations I have
written a fairly detailed specification including error conditions etc., to
demonstrate the level of detail I have in mind. The rest of the operations
are currently just sketched out with examples but are not yet fully
specified (I didn't want to get too carried away before getting wider
input!). I have also left lots of "Editor's Comments" to highlight areas
that I think definitely require further discussion.

Please note I have no intention of landing this full text in a formal group
report, it is only meant as a starting point for discussions, because I
think it's easier to discuss something concrete rather than purely
theoretical. I welcome your comments (inline in the document using the
comments feature, here on the mailing list or on GitHub
<https://github.com/w3c/web-thing-protocol/>), but also potentially other
strawman proposals outlining alternative approaches based on your own
experience if you feel the need.

Following this initial discussion, I propose we create a more formal report
on GitHub using ReSpec, and ensure every paragraph is carefully reviewed
before landing.

I very much look forward to receiving your feedback. Let's get this thing
started!

Best regards

Ben

-- 
Ben Francis
Chair, Web Thing Protocol Community Group