Meeting minutes
Agenda
<cris_> https://
Cristiano: Agenda tbd...
past meeting
<cris_> Nov-28
Cristiano: will chair until Daniel comes
No objections on the past minutes, accepted.
issue 364
<cris_> Issue 364 - Discovery API - Alternatives
Zoltan: one question is how the algorithm for traversing/exploring TDDs defined
Cristiano: some info exists in the Discovery spec, if not a defined algorithm
… in the Discoverer process section
Zoltan: should we define the formal algorithm with dependency to the Discovery spec
Cristiano: is this a Scripting algorithm, or should it follow the Discovery spec?
Zoltan: I think we can define that algorithm and get it reviewed by the Discovery TF.
Cristiano: a lot of MAY clauses in the Discovery spec (mostly suggestions).
Cristiano: we can interpret that to specify a discovery algorithm
Cristiano: I agree we could do later exposing the option of how deep the TDD traversal is done
Jan: agree to leave it for later
Zoltan: a second question in the issue: if we ever do local discovery, will it be a separate method or a dict option
… I suggest separate method
Cristiano: agree about separate method
Daniel: we can decide that later, too
Daniel: maybe we can open an issue to track that
Cristiano: agreed
next calls
Daniel: we have 2 remaining meetings this year, on 12 and 19 Dec. Then on Jan 16.
Cristiano: added to the wiki?
Daniel: yes, in the global wiki
<dape> https://
PR 441
<cris_> PR 441 - Align the discovery API with the Discovery spec. WiP.
Zoltan: a question on whether to use promise for discovery
Daniel: that aligns well with the other conformance classes
Zoltan: we have a single generic discovery algorithm, a meta-algorithm used in all the others
Daniel: can we express conformance classes in Web IDL?
Zoltan: no, only in prose
Zoltan: we might miss the specific directory discovery algorithm, factored out in a separate section
… so please review that with an eye on that
… I could add another commit to include a skeleton
Daniel: by just looking at the Web IDL, cannot tell what belongs to conformance classes
Zoltan: we could do separate API objects for the conformance classes, but usually that means separate spec, because of the security, privacy and permissions.
Zoltan: complex namespace hierarchies would not be good
Cristiano: and not supported any more
Cristiano: so, we need to review the PR and eventually add commits
Daniel: what about multiple URL types for discovery, are they specified in the Discovery spec?
Cristiano: there is a section about CoAP, but it will be addressed maybe in the next charter
… right now we can support CoAP and HTTP
Daniel: maybe the requestThingDescription() might remain open in the end
Zoltan: good point about how are we testing the API? test suite...
Zoltan: I suggest we define a narrowed use case for requestThingDescription(), not a generic do-it-all method
Cristiano: other new features, like cancelable actions are for the next charter
issue 417
<cris_> Issue 417 - emitPropertyChange does not take low-level event apis into account
Cristiano: will discuss it later
issue 409
<cris_> Issue 409 - Harmonize the exposing process
Zoltan: makes sense to revive the discussion, and eventually conclude in this round
adjourned