Meeting minutes
Agenda
Ege: two: minutes review, then coap and binding registry
Minutes Review
Jan 29
https://
Ege: Any comments?
Ege: minutes of wednesday are approved
Jan 30
https://
Ege: Any comments?
Ege: minutes are approved
Binding Templates PRs
Merged PRs
https://
Ege: these are trivial PRs that have been merged
PR 395
<kaz> PR 395 - Fix line endings
Cristiano: I want to discuss the PR 395
… I am not sure why the editorconfig is changed
… it can be kept to be safer?
Klaus: in an ideal case, all computers would use the same line ending
… Linux, Windows and MacOS do it differently
… you can deal with this in different ways, like forcing a specific line endings
… we had issues in this repo as some editors do not support editorconfig
… an alternative is to not force anything, just let people use what their OS does
… gitattributes file will do the conversion for you
… change to your OS in checkout, and in commit convert back
Cristiano: I am familiar with this. Keeping end of line would be safer no?
Cristiano: less bytes wasted
Klaus: safer would be to let git handle it
Cristiano: I cannot remember but there were issues not having
Ege: let's see if we notice an issue
Cristiano: the issue will be only locally
Ege: we can notice that as well
Klaus: only in makefiles this is an issue
Cristiano: I remember that. Also sometimes in docker
Kaz: I personally prefer Linux style but GitHub mechanism is fine
… it would have been better to merge it in the call after the explanation
Ege: you are right. It is a fundamental change in the end
Ege: I could not find the right label as well, let's add something at least
Ege: any further remarks?
Ege: any opinion on the other PRs?
OPC UA URI Scheme
<kaz> Issue 2069 - URL Scheme Decision
Ege: do we know the decision?
Cristiano: We almost settled the discussion last time. I could not join
… we should check with daniel and sebastian about it
Ege: I don't know the final decision's documentation though
Cristiano: it should be the query parameter
Cristiano: there was also a small discussion on using nodeid or just id
CoAP Confirmable Requests
<kaz> PR 392 - feat: introduce cov:confirmable vocabulary term
Jan: klaus made some changes to the PR and they are already approved
… it is about a hint for the consumer to add an option called confirmable to the requests
… it is called cov:confirmable, a boolean
… if you don't use UDP, it is not applicable
… there can be some further guidance for the TD designer on when to use these terms
… it can be a follow up PR
Ege: anything to add?
Klaus: everything looks good
… what triggered this change though? Is it fixing the issue?
Jan: matthias kovatsch brought up the issue in the WoT Week
… knx iot devices were not always turned on.
… however, we should probably work more on this. So that we can know when the device is reachable, i.e. understanding the sleep schedule
… this way, at least the Consumer is not trying to send confirmable requests
Klaus: it would be good to continue the discussion
… and learn more about the background
… if you know that the device wakes up every 5 mins for 20 seconds. We can put that in the TD, which can apply to other protocols
… then you need to know in which part of the schedule we are, synchronize and then send the requests precisely
Jan: I wanted to check if there are existing solutions and algorithms
Klaus: Have you checked with Matthias?
Jan: he was pinged in the issue but maybe I should send an email
Koster: I concur with what Klaus has said
… there is a separation of concerns like Klaus said. Maybe there should be an interface for the Consumer to influence this schedule. Retries create network congestion
… it is more about the configuration of the server
… it should be separated from the normal get/put operations
… most systems won't want you to change it or the system cannot be changed
Klaus: this PR is about a specific use case. The TD creator knows it better that the request is confirmable
… if it makes it better for the people, I don't oppose but we can explore this further
Kaz: The direction is ok and fine but it would be nicer to have a more detailed explanation in the specification text
… like when to use this feature
… so I agree with Klaus on that
<kaz> related Issue 389 - CoAP: Vocabulary for (non-)confirmable messages and retry patterns
Jan: I will try to follow up on this PR
… as Klaus said, it would be valuable to add this to other protocols too
Ege: can you create a user story in the use case repo? we can provide the background in that way and motivate the work
Binding Registry
Ege: I have worked on the PR 378 to make it mergable
PR 378 - Registry Requirements Update
Ege: this is a messy PR because it was collecting opinions
… I have removed the opinions etc. into issues that are linked now
… this way the PR is mergable and we can continue the discussions in the issues for the specific points
Ege: Is it fine for everyone to merge it as is and continue the discussions step by step in the issues and address each issue separately?
Ege: any opinions?
<kaz> (no objections)
Ege: we are fine to merge
… please comment on the specific issues and we can make PRs as well
Ege: anyone want to bring up a topic for discussion?
WoT TD Project Progress Check
<kaz> WoT TD Project
Ege: We are getting more issues than we are handling, seen in the unsorted column of the GH project
… anyone volunteering to take some of the issues here?
Issue 2068
<kaz> TD Issue 2068 - Specify the result of queryaction operation
Cristiano: should once close it at the moment
Ege: we can move it to "Parking"
Cristiano: ok
Issue 2066
TD Issue 2066 - Typo: approache -> the … approach
Cristiano: any PR?
Ege: it is in PR w3c/
Jan: I have experienced that before. I should have opened an issue before
Ege: will check with Daniel about PR 2070
Issue 2071
TD Issue 2071 - json-ld playground Error thing definition ontology/schema
Mahda: you can assign 2071 to me
Ege: we will see each other next week. no cancellations
<kaz> [adjourned]