Meeting minutes
Agenda
Lagally: (goes through the agenda)
… Testfest next week is important
… anything to be added?
Ege: wondering about how to test the Profile spec
McCool: should be discussed during the Testing call later
Lagally: also we have 10 mins for testing prep during this call
Minutes
Lagally: any problems?
(none)
approved
Schedule
Oct 3-6 Plugfest / Testfest for Profile 1.0 Oct 10-15 Incorporate Plugfest/Testfest results, prepare CR draft 2 - weeks review CR transition End Oct
McCool: one correction
… 3-6 Oct for Testfest
… should be Tuesday till Friday
Lagally: meant to be Mon-Thu
(need to be fixed later)
Contributions
PR 266
PR 266 - Refine sync vs. async action protocol binding - closes #259
Ege: Daniel wants to talk about this, but a bit late due to conflicts
Lagally: lets talk about this later then
PR 277
PR 277 - remove http-webhook-profile-protocol-binding-general-2 assertion - closes issue #229
Lagally: would propose we merge this
McCool: we actually need assertion for each MUST
… should be careful about how to handle that
… capitalization for MUST would be important for assertions
… so I agree
Lagally: would merge this then
Ege: ok
merged
PR 295
PR 295 - Prepare Testfest/Plugfest - informative assertion markup - generate manual.csv
Lagally: we have to work on the distinction
… rfc2119-assertion vs rfc2119-informative-assertion
McCool: agree
Lagally: ok
… the point here is that we have an informative section within the WoT Profile 1.0 spec
… also generated manual.csv file for testing
Ben: want to check if all those marked as "rfc2119-informative-assertion" will be removed from the implementation report
… because I think it's too early to remove them from the implementation report
McCool: implementation report defines the assertions
… and we need to provide test results for the assertions
… we should make the decision right away
<Ege> +1 to ben
McCool: the decision is going forward
Ben: I think we need more evidence
… we're still guessing we would not get implementations for those features
McCool: technically, we can include all the features and remove them later, though
Lagally: from my viewpoint, we need to focus on test results during the Testfest next week
… making the SSE section non-normative would make sense
Ben: we can't tell those features could be supported until we actually test them
<Ege> +1 to deferring
McCool: we could do another PR to rebase this
… before finalize the spec
Lagally: if people prefer, I'm OK with deferring it until the Testfest
… what if I exclude the assertions within the assertion.csv file
McCool: we can wait
Lagally: ok
… let's hold this until the Testfest
PR 286
PR 286 - Clarify that href member of ActionStatus can be used by cancelaction - closes #283
Lagally: would like to merge this
merged
PR 287
PR 287 - Consolidate date-time references to RFC 3339 - closes #276
Lagally: changes requested
Ben: note some timestamps which are valid ISO 8601 are not valid under RFC 3339, referencing ISO 8601 is not really enough.
Lagally: need more discussion
PR 288
PR 288 - Remove empty Semantic Annotations section
Lagally: any concerns?
(none)
merged
PR 289
PR 289 - Remove Links section - closes #255
Lagally: kind of big PR
<Ege> +1 to removing everything
McCool: also wondering if we really want to remove all this
Ben: Lagally asked me to apply the changes based on the issue 255
… it's difficult to define "support" here
… we could leave the section, but the features are not implementable
Lagally: you could store the resources
McCool: the question of TD is that it's open-ended
… we should write an assertion for this
Sebastian: good example of Testfest's helping us understand it
<Ege> +1 to sebastian
Sebastian: unfortunately, not really confirmed yet
Lagally: how to test PDF markup as a proper signal?
Sebastian: we should provide some guideline
… not against Profile itself
… but it seems to me we're not ready to test the Profile spec
Lagally: understand your concern
… but we had and have similar problem with TD 1.0/1.1 as well
… we have to discuss what to be done
Ege: PDF is a bad example
… if it can't be handled by some Consumer, can be ignored
McCool: it's still useful for documentation purposes
… Profile is limiting the list of supported types
… making it clear is important here
Ege: the purpose of the Profile spec is making it easier for developers to implement WoT systems
… how to decide the constraint?
Lagally: can you provide counter examples on what to be supported then?
… would suggest we continue the discussion on GitHub
Ben: agree with Ege
… the assertions say "Consumer MUST support..."
… also agree with McCool
… but don't think this section 6.5 Links would add anything for implementations
Kaz: given the discussion so far, would suggest we defer this PR as well until Testfest
… and revisit this in 2 weeks
Lagally: agree
PR 293
PR 293 - Allow security metadata on Forms - closes #292
Lagally: keep it open and ask for further review
PR 294
PR 294 - Make security bootstrapping mandatory - closes #250
Lagally: kind of big changes
… and need further review
PR 266 - revisited
PR 266 - Refine sync vs. async action protocol binding - closes #259
<McCool> (time check - need to end on time, main call)
Lagally: what is the conflict with TD?
Daniel: if I'm a Consumer, would expect to get the boolean within the response as well
Lagally: this PR fixes the problem, doesn't it?
Daniel: only for synchronous
Ben: what it doesn't fix is asynchronous case
… it's still a bit ambiguous within the TD spec
<benfrancis> See https://
Kaz: I had been objecting to add the sync/async options to the Thing Description spec itself, but that was added already
… anyway, Profile and TD should be consistent with each other
… and also we still need to add clarification to TD spec
… we should be aware of those points, though merging this PR itself is OK
Lagally: would merge this PR itself
… and continue the discussion on TD spec, and also consistency check
merged
[adjourned]