WoT vF2F in October, 2021 - Day 1

(There is some trouble with the W3C Web services including the WoT wiki and Mailinglist archive today...)

Agenda

<McCool> https://github.com/w3c/wot/tree/main/PRESENTATIONS/2021-10-online-f2f

McCool: please finalize open day schedules
… I have issues labeled in a certain way and we will discuss these today and reach resolutions
… please upload presentations beforehand to the link above
… many editorial changes from last editor's draft
… we have dealt with anonymous TDs, i.e. TDs with no ids
… for extension to other protocols, we have done JSON based pagination as well. Not clear about long TDs yet
… and also worked on the event api
… we have labeled issues, for example with 2021-10 F2F
… in order to raise awareness for the issues and get input from the group
… Issue 224
… use of @type. should it be mandatory
… it is important for shacl validation or in the future having TMs in the TDDs
… we will discuss in detail later
… also issue 209
… issue 208 for read only TDDs
… issue 190 regarding the id field in the TD field
… is it the Thing id or TD id
… issue 156. what do we do about jsonpath being a draft in ietf
… issue 150 about defining error types

Issue 208

Farshid: currently the spec is for read and write and this is about having a directory that can be only read

McCool's summary

we still need more discussion

Issue 224

TDDs need to deal with missing @type #224

<Zakim> FarshidT, you wanted to react to mjk

Ege: what does a TDD implementor that doesn't implement all the features do?

McCool: its TD will not have all the affordances

Ege: but that means that I need to look at every TDD TD I get, even if it is full compliant

McCool: we can have an array of @type that lists the capabilities of the directory

Farshid: this makes it difficult as well since we can have a type for each different type

Cristiano: I also understand the problem, this would create fragmentation

McCool: unimplemented affordances should not be listed

Cristiano: it should be a group of capabilities

Ege: yes otherwise it is the list of affordances

McCool: Let's reach consensus on this issue though, we have blown our schedule

McCool: I will just comment this on issue since we don't have time for discussion

McCool's comments

Issue 190

Reconsider @type ThingID #190

McCool: ft and ek have different opinions

Ege: my opinion is that in the beginning I remember getting task of changing all ids in the TD spec to dev urn which point to physical devices

Ege: but this is not a strong opinion

Farshid: but the diagram in the spec is for a document not the Thing

McCool: spec is ambiguous, both cases have use cases, we need to clarify the spec

McCool: could you write this in the comment

Andrea: as cristiano said, the ontology is ambiguous

McCool: we should clarify the text and the ontology as well

Kaz: we need to think what we actually meant by "Thing" or "Web Thing" in our specs.

https://w3c.github.io/wot-architecture/#sec-web-thing

[[

Thing or Web Thing

An abstraction of a physical or a virtual entity whose metadata and interfaces are described by a WoT Thing Description, whereas a virtual entity is the composition of one or more Things.

]]

McCool's summary

we still need more discussion including the one on TD

Issue 156

JSONPath and XPath response data models #156

McCool: IETF JSONPath is still not a spec yet. We had some options but now given that all query mechanisms are optional, we can discuss further.
… we need a second xpath implementation for testing

Ege: another option is to write jsonpath spec inside td dir spec

Klaus: it would be difficult to implement all the features. a subset could be interesting

McCool: we need someone to have a look in the spec

Cristiano: I can but what should I look at

McCool's summary

Cristiano volunteers to work on it

Stretch goal issues

Issues labeled as "Stretch goal"

McCool: these are topics that will probably not be in the spec but we can make a last push for some selected ones.

Issue 164

Consider adding "poll hook" to registration (pull keepalive) #164

McCool: Issue 164. Should directory poll the device and check for activity and correctness.

Issue 111

Geolocation query filters for AR #111

Geolocation propoasal

McCool: Issue 111. Static and dynamic use cases.

There is an IEEE std for geolocation queries. It would be better if we wait and then take one of these standards.

McCool: I have a proposal at discovery repository/proposals/geolocation.md that ideally should be a separate note. Maybe a REC track later.

Kaz: I'm OK with this work itself and also splitting this work as a separate GitHub repo, but we should clarify our publication plan for both the original WoT Discovery spec and the possible WoT Geolocation Note given we're getting close to the end of our Charter period.

McCool: let's talk about that during the main call tomorrow

Other issues to be closed

labeled as "2021-10 F2F" and "Propose Closing"

197, 178, 100, 98, 93, 56, 48

Issue 93

Consider using "JSON Lines" for large TDs #93

McCool: can leave this open

Kaz: Farshid mentioned we could close this issue 93 and continue the discussion on 117

McCool: kind of superseded by the new issue

Issue 48

How to avoid duplicates in a directory #48

McCool: we did have problems related to this issue
… probably we should close this issue itself, and then think about more detailed issue

Kaz: we've talked about issues on id already
… so we can close this issue 48 itself
… and continue further discussion on the thread about id

McCool: would like to close this issue 48 itself
… and create another issue on anonymous TD, etc.

Testing

McCool: introduction and explorations
… big focus is directory service fr explorations
… directory service is the largest part of the Discovery spec
… three implementations here

implementations

McCool: LinkSmart, WoT Hive and Logilab TDD

Andrea: tricky with XPath for LinkSmart
… JSON doesn't support XPath itself

McCool: if we don't get sufficient implementations, we can mark the feature as "at risk"
… need to refer to the IETF spec as well

Next meetings

McCool: TD on Thursday, Oct 7
… Sebastian will chair it

Sebastian: yes
… including the plan for 2.0

McCool: note that we still have trouble with wiki, etc.
… please upload your slides to the GitHub

PRESENTATIONS area for Oct vF2F

Other WoT meetings this week

McCool: we'll have the main call tomorrow
… no TD call
… conflict with the Architecture call, so no Architecture call
… I've updated the WoT main wiki about the cancellation, though it's not accessible now
… see you all tomorrow at the main call!

[adjourned]