– DRAFT –
WoT-WG - TD-TF

Minutes

<kaz> May-25

ege: review 05-25; old, many cancellations, testfest...
… discussion of improving SVG figures
… discussion of deriving ids from TMs
… please expand ca into a full name
… JSON version 7; vs validation; discussion in profiles

ege: documents themselves are fairly clear

ege: fixed a compatibility problem, issue 1510
… finally, had a resolution to publish a CR candidate

mm: suggest we also update the schedule today, now there is a CR candidate in flight

ege: any comments on the minutes?
… no objections, let's publish.

Binding Templates

updates

ege: now have netify previews, so not only root gets rendered
… but needs coc and license document

mm: note that repo has automation now

ege: also need to run a script...

<Ege> https://github.com/w3c/wot-binding-templates/issues/163

mm: perhaps some instructions on how to update

ege: this PR itself is just the coc and license for netlify; and this is automatic

kaz: agree with McCool, and we should document how to use it

ege: created issue #163 to follow up on documentation

MQTT Binding, PR #156

<Ege> PR 156 - Improve the mqtt protocol binding template

<kaz> preview

ege: think we need to talk about the use of retain
… what does it mean, it to be set, or IS set?

cris: maybe better to not include, is a side effect of using readproperty
… but maybe have to consider how to deal with short-lived events
… this was one reason I added this; might be some corner cases where it is not needed

ege: may also be cases where published messages to not change value
… maybe need a chapter that explains this
… example needs to be improved, show two forms
… or three
… to show the different cases
… putting multiple operations in the same form is tricky

cris: probably better to have different forms

ege: and we should recommend people separate them to avoid ambiguity

ege: another issue is using href in combination with topic and filter

cris: also some formatting issues due to changes in ontology file
… anyway, I added these fields to follow spec for MQTT closely
… then it becomes clear where topics and filters go
… can you be specific about what should be changed?

ege: I think just more information about where topic and filter apply
… needs to be more explicit
… e.g. in subscription, should not include topic, etc.

cris: and URL scheme is just for broker

ege: and say URL is used for connection request

kaz: question, main purpose of binding template, main document
… main doc lists set of protocols
… we should remember the Profile spec handles https as a "core" protocol, others as "bindings"

ege: binding is not about one to another, is how to use one
… so not about binding mqtt and http, but using mqtt in TDs

kaz: I know that point, but the current description within the binding temp overview is not enough

ege: agree, I am working on that

<Mizushima> MQTT Binding Template

<Mizushima> +1 kaz

mm: I also find it odd that http is built-in to TD but others are extensions

ege: ideally next TD 2.0 everything will be in an extension

mm: also, minor point, avoid refs in abstract, can be taken out of context, then refs will be broken

cris: sure
… was in older version, not really being dealt with right now

ege: (eg captures some notes on PR #156)

ege: looking ahead a bit, we have a few other PRs on CoAP, etc.
… would be good to be consistent, use the v suffix, etc.

ege: don't like it, but it avoids confusion with the protocol name
… should use cov, not coap; otherwise prefix looks like a URI scheme

cris: also should say that in case of collisions, longer prefixes can be used.

ege: (captures notes in issue 120)

suptopic: CoAP issues

ege: discussion of composite things, multicasting in CoAP
… can be used to for, example, turn off all lamps

mm: multiple issues here; also security, discovery
… but I think core issue is a "group affordance", URI scheme...

ege: for future work, just want to table

TD

Update

<Ege> https://github.com/w3c/wot-thing-description/issues?q=is%3Aopen+is%3Aissue+no%3Alabel

ege: in github, is feature to look at unlabelled issue; I went through issue and added some labels
… now have an action to add a "needs triage" label automatically if there is not a label

<Ege> https://github.com/w3c/wot-thing-description/issues?q=is%3Aissue+is%3Aopen+sort%3Aupdated-desc

ege: can see some old issues getting updated

ege: also identified a lot of issues that can be closed
… probably

mm: could send an email to close in a week

ege: or just close and people can re-open

kaz: we need to do work on CR-cand static draft

mm: to be clear, I am not working on this, only on IR

PR 1532

<Ege> Allow placeholder for required and enum in TMs

ege: is fix to JSON schemas for TM for required and enum
… enum was not allowing placeholders in TM
… main thing that was updated was the TM

ege: (merges after fixing conflict)

PR #1517

ege: Daniel not here, will skip for now

kaz: title of this should be "Create CR Candidate WD"

ege: ok, changed

PR #1501

<kaz> PR 1501 - WIP: Add at-risk hints

ege: todo.csv takes place of this, and also does at-risk markup, so can close without merging

mm: note however we do need to explictly document at-risk item in sotd section
… but can do in a couple of weeks after we clean up the todos

extra asserts

<kaz> Issue 1527 - [Pipeline] Extra-asserts and depends.csv should be automatically generated

mm: ideally, but hard; suggest we do it manually for now

<kaz> Ege's comments

Issue 1511

<kaz> Issue 1511 - Restructuring td-json-open assertions

mm: suggest we reach out to internationalization people
… is redundant, so how can we clean up to make it easier to manage testing while still acheiving objectives

issue 1439

<kaz> Issue 1439 - OAuth2 Assertion: other-flow assertion is removed from spec but visible in implementation report

ege: oauth-other-flows, but seems to be gone now

mm: I think I fixed this when revising IR

ege: ok, let's close

issue 733

<kaz> Issue 733 - [At-Risk] Implementations of PSKSecurityScheme needed

ege: very old, but still open
… also 732, 735; same issue
… closing all

<kaz> Issue 732 - [At-Risk] Implementations of CertSecurityScheme needed

<kaz> Issue 735 - [At-Risk] Implementations of PublicSecurityScheme needed

issue 888

<kaz> Issue 888 - [question][behavior][testing] Should writing to an observable property trigger new values being pushed

ege: does writing a property trigger an observe update?

mm: TDs are descriptive, if impl does not do what the op says it does there should be a different op
… but the definition of op should perhaps not say "when", which implies immediately, but should add a note that some implementations may defer or queue updates...

ege: let me take a note about this

mm: we *could* add an op for "strictobserve" but could also use an event if you want something different than eventual consistency, which is what observe is for

CR transition

ege: labelled a number of issues that are needed for the CR transition

ege: easy one is missing keywords in assertions

Issue 1523

<kaz> Issue 1523 - Missing RFC-2119 Keywords in Assertions

<kaz> Ege's comments (use SHOULD and MUST)

1519

<kaz> Issue 1519 - ReSpec Error - Form definition

ege: use of "Form"
… this used to be defined somewhere; but also in HTML spec

mm: suggest link to Form section of TD spec for this

<kaz> Ege's comments

Issue 1520

<kaz> Issue 1520 - ReSpec Error - Reference "[ACE-OAuth]" not found.

mm: probably should be RFC 9200; was published as an actual RFC
… not in specref yet, will have to make a local definition

<kaz> Ege's comments

issue 1518

ege: respec bugs...

<kaz> Issue 1518 - ReSpec Warnings - Normative reference defined in informative document

ege: so we can normatively ref an informative spec, but respec complains

mm: we agreed to do this, so issue is how to get respec to stop bugging us about it

kaz: suggest we talk to PLH and Systeam

issue 1343

<Ege> Issue 1343 - Dereference 1.1 context file via new TD 1.1 IRI

ege: think it has been solved already

mm: any instances in the spec of the old URL?

ege: let me check...
… does seem to be use of 2019 in some of the URLs
… I think the current issue is fine, will close, but will open another one about the 2019 URLs

issue 953

<kaz> Issue 1540 - 2019 URLs showing in namespaces #1540

<kaz> Issue 953 - For OAuth2 device flow, should we define a "device authorization" element?

mm: this has been resolved; we are using "authorization" for "device"

ege: closing

issue 926

<kaz> Issue 926 - Add OAuth2 client and device flows

mm: also done

ege: closing

issue 1243

<kaz> Issue 1243 - Definition of Backwards Compatibility

ege: defn of backward compat

ege: have added some clarification in text regarding validation, and now all old TDs pass

ege: closing

issue 854

<kaz> Issue 854 - Clarify read-only/write-only behavior

ege: about read-only vs. write-only behavior

mm: would like to see change to readable and writeable, but not feasible now
… issue though is about default value for readOnly, etc.

ege: will create issue to look at this in TD 2.0

mm: and also we should complain again to people working on JSON schema RFC

<Ege> https://github.com/w3c/wot-thing-description/issues?q=is%3Aopen+is%3Aissue+label%3AV1.1

label v1.1

ege: lots of open issues

mm: anything related to canonicalization should be closed, we can reopen for TD 2.0

ege: ok, will do offline

issue 1396

<kaz> Issue 1396 - Complete TAG/Security Wide Review Request

mm: TAG review *request* is done, we can close this issue

combo minCount 2

<kaz> Issue 1438 - ComboSecurityScheme: oneOf and allOf can be strings according to vocab table

mm: this is in ontology, not getting caught by render script
… in ttl, is in minCount 2
… possibly JSON schema does not include this constraint

ege: added note to issue

wrapup

ege: will close all proposed closing issues and send an email
… and will also close canonicalization issues as discussed

<kaz> [adjourned]