– DRAFT –
WoT Discovery

Minutes

<kaz> Mar-20

<kaz> approved

Issues

<McCool> https://github.com/w3c/wot-discovery/labels/resolve%20by%20PR

<McCool> https://github.com/w3c/wot-discovery/labels/Testing

McCool: So we have two labels, "Resolve by PR" and "Testing", need to check that we have not missed any issues
… "Resolve by PR" issues might also be resolved early (in a "won't fix" manner)
… does somebody here volunteer to address some of these issues?

Andrea: I can deal with some of the technical(?) issues

Issue 417

<kaz> Issue 417 - Should we make SSE informative?

McCool: Regarding SSE (issue 417), I am not sure what the current problem/state was
… is referring to the Events API
… one issue here is that we did not mark SSE as being as risk
… even though we said so in the issue itself
… things like the diff query assertion are minor
… other assertions are not at risk, we have implementations, so we do not need to mark the whole feature as at-risk
… some of the assertions might not need to be MUSTs in environments like LANs, so we might need to revisit these later
… (adds a comment to the issue)
… one thing we can do is moving this to Discovery 2.0 and then discuss alternative mechanisms such as WebHooks

<McCool> https://github.com/w3c/wot-discovery/issues/417#top

McCool: I will close this issue with comment and then create a new one regarding a scalable event mechanism
… (creates issue 468)

<McCool> w3c/wot-discovery#468

McCool: issue 417 has been linked automatically to this new issue

McCool: Now let's get back to other "Resolve by PR" issues

Issue 64

<kaz> Issue 64 - XPath queries of vocabulary using namespaces

McCool: XPath is still not a normative specification, so we can also label this issue as "informative"
… (adds the label)
… one thing we could do here is defer it to Discovery 2.0
… when XPath could become normative
… a bit unclear what to do here
… does anyone want to volunteer to work on XPath?

No response to the question.

Issue 91

<kaz> Issue 91 - Clarify relationship of discovery to lifecycle

McCool: This issue is dealing with the "relationship of discovery to lifecycle"
… reading through this, I essentially proposed doing nothing
… it's clear that discovery is not the same as onboarding
… should therefore be informative or moved to Architecture
… would rather say nothing than saying something wrong
… does anyone feel different?
… I would propose deferring this to Discovery 2.0
… (adds a comment and the respective label, removes the "Resolve by PR" label)

Issue 69

<kaz> Issue 69 - Add sequence diagrams

McCool: This issue is about sequence diagrams
… I guess I can do that myself

Issue 439

<kaz> Issue 439 - Use Case due diligence

McCool: So we essentially made an assessment, is an ongoing process
… filled out the blanks in the table
… can be considered done
… (adds a comment)

<McCool> https://github.com/w3c/wot-discovery/issues/439#top

McCool: any objections to closing this issue?
… hearing no objections, closing it

Issue 466

<kaz> Issue 466 - Removing the RFC 9175 assertion?

Jan: Is about the assertion regarding the mitigation of Denial of Service Attacks in the context of CoAP. I think the wrong document is cited. It is at risk anyway, though

McCool: We could make a PR and make the statement informative
… we are allowed to drop assertions and turn them into informative statements

Jan: Citation should be adjusted, RFC 9175 instead of an Internet-Draft by the IRTF is cited

McCool: (adds a comment with a proposal for a reformulation)

Jan: I can volunteer to do a PR

McCool: (assigns Jan to the issue)

Issue 465

<kaz> Issue 465 - TDD Thing Model is missing contentType in response objects

McCool: Also opened by Jan

Jan: Deals with the issue that you have to define a contentType in the response objects

McCool: Since the content of the response is empty, the contentType does not really matter
… need to add an informative note, since the issue in the TD repository is marked as "Defer to 2.0"
… (adds a comment with a proposal)
… application/octet-stream can be used as a fallback

Jan: I can do a PR

McCool: Just make it an informative statement, not an Editor's Note
… (adds a new label "Resolution Proposed")
… (also adds the label to previously discussed issues)

Issue 463

<kaz> Issue 463 - Align Context URLs with TD specification

McCool: We never finalized the discussion regarding @context URLs
… there is a PR against the DID repo
… where a @context URL with a year is used
… with /did at the end and discovery inbetween
… therefore there is the discussion to use an "ns" URL instead
… the TD specification also uses a year in its @context URL
… is anyone aware of any kind discussion regarding namespaces in the context of the TD specification?
… (searches the TD issue tracker)
… (does not find a related issue)
… seems as if changing the namespace has not been considered so far
… the PR in the DID repo has not been merged yet
… they are waiting for us to have a deferenceable URL in the PR before merging it
… so we can still update it
… there is also Issue 148 which is fixing it and also uses "ns"
… let's see if we can make a decision here
… (adds a comment to Issue 463)
… I propose using also a different namespace for the TDD vocabulary to avoid having problems

McCool: As a resolution, I propose using 2022 in the namespace
… is a normative change, but should be editorial
… however, implementations need to be updated
… in order to re-validate the assertions
… I will do a PR
… does someone want to help with the ontology file?
… will not change anything just yet, will wait for feedback

<kaz> McCool's comments

Dev Meeting Preparation

<kaz> Slides on TD

<kaz> Slides on Architecture

McCool: We made some preparations
… prepared slides walking through assertions
… meeting on the 30th
… TD slides are done, currently preparing Architecture
… should do something similar for Discovery
… I can do a first draft
… currently waiting for stabilization of the opening sections
… then I can create something similar for Discovery
… I put the references in the Wiki, you can have a look through them

[adjourned]