Meeting minutes
Minutes Review
<kaz> Feb-19
Cristiano: I checked the minutes and they were very useful as a short summary for the last time
… I encourage everyone to do so as well
… so as usual, we discussed the PRs
… we are going to revisit the one first one today
… very important is that we assigned Daniel to label the issues relevant for the TD taskforce
… that's it, everyone okay with approving them?
… there are a couple of IRC messages in-between the regular minute-taking, maybe we should remove them
Kaz: Will do
… fixed, please reload
Cristiano: If there are no other points, minutes are approved
Minutes are approved
Updates
Cristiano: As you probably heard, we are on a bi-weekly schedule
… if you feel like there is there the need to change the schedule again, just let us know
PRs
PR 489
<cris_> PR 489 - Better types for Scripting API
Cristiano: This one is basically stalled, not sure if it can be fixed, for now this is mostly a reminder for us
PR 534
<cris_> PR 534 - fix(InteractionOutput): don't require schema.type in value function
Cristiano: I want to talk about this one since Jan pushed some commits
… I remember that we discussed this last time and that there were some things that needed to be fixed
Jan: I addressed the comments
… I also discovered that default keyword (adressed to be a candidate for the algorithm)
… it is just an annotation only for documentation
… it is from json schema specification
… and we can simply skip it
… other than that I think the PR is ready to be merged
… we can check the rendered document
Cristiano: I share your view on default
… was only meant as a documentation for developers
… did not find the reference to JSON schema, glad that you did
… we might be able to put some additional semantics in Web of Things
… but this not well-defined yet so I am definitely fine with keeping it as is
… (opens the rendered diff)
<kaz> (Regarding the diff, Kaz has no problem with Vivaldi browser, but Jan has problem with Firefox and the diff is highlighted by "black" :( )
Cristiano: from my point of view these are fine
… maybe we need to wait for Zoltan
… is an improvement for developers, but not a perfect algorithm yet, can be further improved in the future
Daniel: I approved the PR a while ago
… gave thumbs up to the latest changes
… but don't know how to proceed with the whole data schema topic
… needs some further improvement in the future, but is an improvement already
Cristiano: Then we can go ahead with merging this PR
… (adds a comment to the PR)
… if no one objects, then we can merge it
… of course we can always check back with Zoltan and fix it later on
No objections, PR is merged
PR 545
<kaz> PR 545 - feat: expand exploreDirectory algorithm
Cristiano: This is another draft PR by Jan
… previously discussed this
Jan: I tried to expand the exploreDirectory algorithm
… maybe we can cover also the query api
… it's an early status
… I will work on it for the next meeting
… I want to add the ability to filter things
… basically using the search API
Cristiano: do you mind to share some early ideas
Jan: at the moment we can only use the things property
… to retrieve thing descriptions
… that's the only way implemented in the Scripting API spec
… I wanted to add also actions for search things
… search API
… with that we can use json path and sparql
… maybe we can have a wrapper algorithm
… to identify which endpoint to use
… we can also fallback to things API if search api is supported.
Cristiano: do you want to create an abstraction layer for queries?
… or simply sending string queries?
Jan: just strings added to a basic payload that have an hint to the runtime about with query language to use
dape: I will keep it simply and it matches very well with the API
Cristiano: +1
Issues
Issues relevant for the TD Spec
<kaz> Issues on wot-thing-description repo related to Scripting API
Cristiano: Maybe you can present the list you've got so far, Daniel?
<JKRhb> i|dape:I will keep|scribenick: JKRhb|
Daniel: Sometimes there are several places we can link, but generally have labels defined both in the TD repo and the Scripting API repo
… we need more feedback on how to handle actions, especially when it comes to querying and cancelling
… this is currently underspecified in the TD spec
… others are slightly less important, for example canonicalizations
… was postponed a bit because people said that JSON-LD offers this functionality, but not sure about that
… another issue is related to properties that are writable but might not process a value exactly as it is put it (e.g., with regard to precision)
… a concrete example would be Phillips Hue, where the value that is actually written is returned
… these are just examples, in general the linking is in the process which then need to be discussed in the TD taskforce
Kaz: Very important work to clarifiy the relationship between TD and Scripting API, thank you for that
<cris_> +1
Kaz: however, the label in the TD repo should also say "Needed by Scripting API"
Daniel: Forgot to talk about this with Ege, will talk to him
… that we have another label, will create an issue for that
… in the TD repo
Cristiano: Definitely agree with that
… there should be different labels per taskforce
… is there anything concrete to help the TD taskforce to move over with these issues?
Daniel: Currently, the TD taskforce is still in the process of preparing its work, no technical work is happening at the moment
… maybe we could pick out a couple of issues that are the most important to us (one or two)
… should form a consensus and then drive the issues forward in the TD taskforce and give our input in the calls
… last week I couldn't join, but I think the technical work has not started, right?
Cristiano: Correct
… we probably also need to provide use cases
Daniel: Some are labelled with the use-cases label, maybe we need to talk to Ege and Michael Koster
Kaz: This work between taskforces should be clarified in general
… maybe the use-case discussion is also related to that
… maybe we need to think about how to collaborate across taskforces in general
… should be part of the WG procedure
Cristiano: Agree, in general we should know how to process issues
… currently, we are kind of making our own rules, should be clarified
Daniel: There are a lot of issues related to use cases, about 100
Cristiano: Maybe we can integrate that in the use cases workflow, using the project
AOB
Cristiano: I think we can close early, any other business?
Semantic API
Daniel: I have something in my backlog regarding semantic API, but I will need to work a bit more on that, will present something at some point
Cristiano: Do you have a use case in Siemens?
Daniel: So far, there is not much happening regarding semantic annotations in general, improving on that could be a great selling point for WoT
… will keep you posted
Versioning
Cristiano: Another thing that came to my mind was the versioning problem
… that could also be put on the agenda for the TD call
<cris_> w3c/
Cristiano: then we can publish a 1.0 version
<dape> w3c/
Cristiano: this is related to issue 488
Daniel: There is a PR (1969) in the TD repo
… but it still a draft
Cristiano: I bring this up because we could then publish 1.0 under a tag and then make a clean up
… will try to talk to Ege to clarify this
Daniel: Agree, versioning scheme should be clear, also to be able to publish snapshots inbetween
Cristiano: Will talk to Ege as a homework
… see you in two weeks
[adjourned]