Meeting minutes
Minutes
<kaz> Minutes from the meeting on Oct 9, 2024
<benfrancis> Minutes approved
Ben: any objections on the last minutes?
<no>
WoT Profiles 2025 Planning
<Ben shows a presentation>
Definition
<kaz> [slide 2]
<benfrancis> "A [profile is a] technical specification which provides a set of assertions such that any Consumer which conforms with the those assertions is out-of-the-box interoperable with any Thing which also conforms with those assertions."
Ben: Profile definition from WoT Architecture 1.1:
… "A [profile is a] technical specification which provides a set of assertions such that any Consumer which conforms with the those assertions is out-of-the-box interoperable with any Thing which also conforms with those assertions.”
Questions
<kaz> [slide 3]
Ben: There are questions:
… should we continue to pursue the existing WoT Profiles 1.0 specification along the recommendation track
… or
… publish WoT Profiles 1.0 as a non-normative Working Group Note and start to work on use cases & requirements for WoT Profiles 2.0 instead?
Issues
<kaz> [slide 4]
Ben: Issues with WoT Profile 1.0
… extends as well as constrains what is possible with a Thing Description alone
… is limited by the features of Thing Description 1.x
… not enough implementations
Ben: are there more issues?
Daniel: what is written covers so far the most
… if the Profile is a subset of a TD then everything is ok
Sebastian: Have the same view as Daniel. Express this as a wish. If go for a Profile specification, should be designed in such a way that it is not confusing.
… Specification documents TD 1.1, bindings etc. Should work smoothly with each other
… Shouldn't have conflicting definitions or features in the Profile which only work with a Profile definition but not core TD specification.
… Concern Currently TD assumes everything specified in the TD document, no assumptions about the payload description for example. There are currently assumptions, e.g. about async actions.
… For user of Thing Descriptions, getting confused. If expecting a specific payload structure, not in the TD. Should not conflict with idea or principle of TD.
Ben: there are already assumptions in the TD because there are and there are also default binding as well as default payload binding including default methods
… profile defines separate set of assumtions, I think that where the problem lies
David: we hoping with Profile how subscribing implementations works. This would help developer. Current version of Profile explains server side event implementations.
… Conexxus does not make implementations but create standards that our implementer members can take
Ben: you have great use cases for Profiles
… we need to manage all the different concepts like binding, semantic contexts, thing models, etc.
Kaz: agree we all comments so far. personally, don't think Profile should "extend" TD itself, but should concentrate on describing "how to use TD for actual applications", but would like to hear from you a bit more.
Profiles 2.0
<kaz> [slide 7]
<Ben shows a Profile 2.0 Strawman Proposal>
Ben: A simpler and more structured approach
… move content of protocol binding sections into Binding Templates as defaults, referenced from Profiles
… core specification which defines the profiling mechanism and a registry of profiles
<benfrancis> Ben's message on rebooting Profile
<benfrancis> Strawman Proposal for "HTTP Protocol Binding 2.0"
Ben: individual profile documents constrain specific extension points of Thing Descriptions
<benfrancis> Strawman Proposal for "HTTP Basic Profile 2.0"
Preservere or Pivot
<kaz> [slide 8]
Sebastian: Regarding 2.0 strawman proposal. What I see so far I quite like. Move out the protocol related topics, put them in protocol binding template. Profile documents only pick the definitions/building blocks. Explain what they constrain. Not examined in detail yet, right direction.
Ben: Pros for a 1.0 REC
… Interoperability for WoT 1.x specifications
Ben: cons ...
… Either extends other WoT 1.x specifications or is very limited in features
… May not be worth publishing if it doesn’t go much beyond HTTP binding in TD spec
Ben: pros publishing as 1.0 Note
… Don’t waste time refining and publishing something that may not get implemented
… Profiles 2.0 could work better with Binding Templates and benefit from TD 2.0 features
Sebastian: Personal opinion is that we go with a 1.0 Note publication because it simplifies everything regarding publication. Already a lot of material and content available, would avoid the requirement of implementations. Would still need to clarify topics we raised at the beginning of the meeting, would like to work more on 2.0 which is aligned with TD 2.0.
… Would prefer to go with Note for now, quick publication clarifying remaining points. Go for 2.0. Question: What happens to 1.0? Deprecated.
Luca: We have no consenus on 1.0 REC publication, support Note publication
… regarding implementation I wrote the 2nd implementation
… main concern is the paper work
… I would just focus on getting everything right for 2.0 REC
Kaz: Agree with both. In addition, "interoperability" for WoT specs should be guaranteed by the WoT specs themselves like TD. So let's go for 2.0
Daniel: 2.0 should work smoothly together with what we re going to create in the future. We need also someone that is actively working on the TD side and on the Binding side. Sync is needed closely
Ben: agree about this
Mizushima: I agree with the comments.
… if we not have everything for REC we should make a Group Note
Proposed resolution
<kaz> [slide 9]
Ben: it seems that we have consensus. We can send a respolution to the email list and approve the resolution in the next main call
<kaz> WoT WG Charter
<benfrancis> Proposed resolution: The WoT Profiles Task Force will aim to publish WoT Profiles 1.0 as a Working Group Note by July 2025 and write Use Cases & Requirements for WoT Profiles 2.0 by October 2025
<sebastian> +1
<Cris> +1
<luca_barbato> +1
<dezell> +1
Ben: any comments or objections?
<kaz> +1
<dape> +1
<Mizushima> +1
<jthomas> +1
David: just want to say thank you. I think this is a great work
SK: we have 3 months, seems ok for the Note publication
Ben: suggest to have meetings every 2weeks
Daniel: are there any pressing issue that need to be clarify?
Ben: there are some things that I wasn't completely happy about webhooks
next steps
Ben: Propose our resolution by email today with a view to passing the resolution in the WoT Main call next week
… hold a meeting next week to triage issues and create a more detailed plan
<dezell> +1
Ben: propose meetings every two weeks after that, with most collaboration happening asynchronously on GitHub
Ben: is it ok to have the call every 2weeks?
<kaz> kaz: OK with the 2nd point and the 3rd point. However, regarding the 1st point, I'd suggest we make our own TF-wide resolution during this call now given we have 10 people here agreeing to that :)
<benfrancis> Proposed resolution: The WoT Profiles Task Force will aim to publish WoT Profiles 1.0 as a Working Group Note by July 2025 and write Use Cases & Requirements for WoT Profiles 2.0 by October 2025
RESOLUTION: The WoT Profiles Task Force will aim to publish WoT Profiles 1.0 as a Working Group Note by July 2025 and write Use Cases & Requirements for WoT Profiles 2.0 by October 2025
Ben: 2.0 work will have close collaboaration with Use Case, TD and Binding TF
Ben: any other comments?
Kaz: if needed we can extend the WG periodbased on the need and our progress in 3 months.
Ben: there is a WoT Community Discord channel. Link is provided at the wiki
<kaz> [adjourned]