– DRAFT –
WoT Discovery

Minutes

Sep-5

McCool: (skims the minutes)

approved

Logistics

McCool: Discovery meeting cancelled next week due to Testfest?
… any objections?

(none)

RESOLUTION: Discovery call cancelled on Oct 3 due to the Testfest

PRs

PR 411

PR 411 - Replace CoRE RD draft with RFC 9176

McCool: any objections to merge this?

(none)

merged

PR 412

PR 412 - Sept 2022 Implementation Report

McCool: bunch of assertions added
… typo fix as well
… some of the assertions moved away
… terminology clean up

McCool: did changes for at-risk features
… would like to merge this
… any objections?

(none)

merged

Updated implementation report

Automatic tests

McCool: shows the updated implementation report

Discovery Implementation Report - 8.1 Automated Validation Results

McCool: Toumura-san, any results on UDP handling?

Toumura: no, not so far

(discussion on tdd-absolute-time)

Farshid: defined by RFC 3339

McCool: not really sure what "interpreted" here

Whenever an absolute time is expressed using dateTime it MUST be
interpreted as date-time as specified in [[RFC3339]].

McCool: maybe the wording need to be fixed

Farshid: it's really implied

McCool: it's just an editorial issue in English
… let me think about that
… (creates a new issue about that)

Issue 413 - Improve wording of tdd-absolute-time assertion

(discussion on tdd-http-unsupported-feature)

McCool: manual test?
… (creates a new issue)

Issue 414 - make tdd-http-unsupported-feature manual

(discussion on tdd-things-create-anonymous-td)

Issue 415 - confirm test results for tdd-things-create-anonymous-td

Manual tests

8.2 Manual validation results

McCool: maybe some of the manual tests should be informative
… There are assertions on "Discoverer"
… we have node-wot and NodeRED as implementations
… some of the results are strange
… e.g., not sure why discoverer-any-order doesn't pass
… need to look at the actual CSV files again
… (looks at the CSV files)
… maybe need to clean up the results

Christian: seems some overlaps there

McCool: yeah, need to clean up logilab results up
… (creates a new issue)

Issue 416 - clean up logilab results

McCool: (assigns it to Christian)

(tdd-context-injection)

McCool: it's weird...
… maybe need bug fixes

An Enriched TD MUST contain in its @context the URI https://w3c.github.io/wot-discovery/context/discovery-context.jsonld.

McCool: the assertion the github.io URL
… that's also weird
… should resolve issue 148

Issue 148 - Discovery context and namespace under w3.org/ns

McCool: we've agreed to go for the option 2 here

2. Continue to use github for now (makes testing easier), add an ednote
saying it will be changed in the final version. (Late binding and back-patching)

McCool: changing the label of the issue 148 to "Resolve by PR"

(tdd-http-representation-alternate-output)

McCool: marking as at-risk at the moment

Farshid: related issue 185

Issue 185 - OAuth2 and SSE Notificiations

Farshid: Andrea might have an implementation

ac: also implemented SSE
… the question is generating the client from the scratch

McCool: right

Farshid: libraries don't provide storage to you

McCool: it's a trade on the performance
… your option 4 looks good but require extra implementation

4. Same as above, but with OAuth 2.0 Token Introspection
* The server checks the token with auth-server periodically and
  disconnects the client if access is revoked.
* Very secure Most resource-friendly approach for the clients Depending
  on the introspection interval, it could put high pressure on server
  and auth-server. It requires some extra implementation on the
  server-side.

McCool: there is an implementation but we don't think that's enough to make this feature normative
… assertions are all about the server, not the client; so we don't actually have tests that check for client-server interop.

McCool's comments

McCool: (creates a new issue)

Issue 417 - Should we make SSE informative?

McCool: can we transition to PR if we can't get enough implementations for this?

kaz: no, need to identify this as at-risk
… or we have to publish an updated draft as a CR

McCool: ok
… we need to continue to work
… but due to the Testfest next week, maybe in 2 weeks
… (adds a note "McCool will prep for mtg in two weeks" to the agenda)

[adjourned]