update Section 8.3

[sebastiankb]

addresses issue w3c/wot-architecture#639

• provide reference to URI scheme RFC3986
• note about IANA

---

Preview | Diff

[benfrancis]

I understand that this change is intended to make it possible to describe existing IoT protocols like MQTT and Modbus in a Thing Description, and whilst I don't personally agree with that I accept that others want to do it.

My concern with this wording though is that because the terms "defacto" and "established" are so open to interpretation it essentially introduces a loophole which can allow any protocol to be a WoT protocol, including non-Internet protocols. In fact it is so open that it arguably makes the assertion meaningless.

If I've understood correctly then someone could theoretically make up their own unofficial URI scheme for a non-IP protocol like Zigbee or Z-Wave and then create a protocol binding template which uses that URI scheme. This blows open the number of protocols that a WoT Consumer might be expected to support even further and removes the compromise we reached to ensure that this couldn't happen.

I'm afraid I don't have an alternative proposal which can work for the MQTT and Modbus use cases but not for the Zigbee and Z-Wave use cases (other than getting their standards bodies to register official URI schemes with IANA), but I'm interested to hear others' views on this.

[sebastiankb]

from today's TD call:

• not going to merge. Some improvements will be done based on @benfrancis comment above
• @egekorkan will provide also the location in the binding spec which also address this aspect --> this link will be also covered by the note then

[egekorkan]

I sort of forgot what I exactly had to do but I have added the link for the relevant part in the binding spec

[mlagally]

This language is difficult to understand, since it refers the entire binding specification. The implication on the URI scene should be clarified.

[egekorkan]

It shouldn't be the case, for me the relative part is shown when I click on https://w3c.github.io/wot-binding-templates/#creating-a-new-protocol-binding

[sebastiankb]

from today's TD call:

• decided to merge
• we need to double check with @mlagally what in particular should be improved based on the comment above --> another PR can take this into account then
• another discussion was about if the section 8.3.1. should be moved to the binding document. there is a discussion in issue Where to place the HTTP Default Binding and which format #1274
• in general we should decide what should be normative in TD regarding protocol binding topics.

[mlagally]

This language contains a reference to the binding template spec in a RFC assertion.
The sentence needs to be split and move the binding templates out of the normative part.