Organize Deliverables for Protocol Bindings

[mmccool]

• Informative deliverables (Notes) for individual bindings
• Mention use of "hub" document as a registry

Owner: Ege

[egekorkan]

#31 fixes this to a degree. @mmccool is the intention here that I define 2 new work items with the bullet points above?

[mmccool]

@EGE, no, I don't think these are work items, just need to mention the Notes under informative deliverables and perhaps edit the "normative" deliverable to refer them. However, let's discuss in the call today.

[mmccool]

But yes, #31 does consolidate the "protocol bindings" and "payload bindings" work items, which is half the cleanup/consolidation needed, I think.

[egekorkan]

My comment from Call of 15.02:

• We should not have REC for each binding, it means having too many requirements in the charter, raising issues in different levels
• We should have a REC for the core document.
• Each binding is a W3C WG Note but the Core Document (which is a REC), specifies additional requirements on top of the basic W3C requirements for publishing notes. These can include having at least two implementations or other requirements we put.

[sebastiankb]

in today's TD call, we did some discussion on the bindings for the next charter. Please find here the slides that I have presented. I like to share the slides in tomorrow's charter meeting as well: WoT_Binding_2.0.pdf

[benfrancis]

@sebastiankb wrote:

Please find here the slides that I have presented. I like to share the slides in tomorrow's charter meeting as well: WoT_Binding_2.0.pdf

This is interesting, thank you for sharing. I was disappointed that this topic didn't get discussed on the call.

I agree in principle with the idea that the top level document for bindings should be a REC containing registry sections and that those registries could point to W3C Notes (to reduce publication overhead). However, I do wonder about the implications of the binding documents being non-normative Notes. As I understand it, W3C Group Notes are really intended for supporting documentation such as use cases and requirements, design principles and best practices - documents that are not intended to be a formal standard.

If the binding documents are going to contain implementable assertions essential for consuming Things and are going to be cited by other RECs, is that enough? Particularly if normative sections of the Thing Description specification like security mechanisms are moved to binding documents as proposed in the draft charter, and if Profiles reference binding documents instead of defining their own protocol bindings as proposed in w3c/wot-profile#285 (comment).

Regarding the naming, I have concerns about calling the documents "protocol bindings" rather than "protocol binding templates", which I have explained in detail in #14. For what it's worth, if the binding documents are renamed from "binding template" to "binding" then I think the top level document should just be called "WoT Bindings" since its main purpose is to be a registry of bindings and define the criteria for being included in that registry, not just to provide a template.

P.S. I'm curious, what is the idea of the "WoT Profile Basic Binding" example in the slides?