Meeting minutes
Scribes for today
scribes so far: Ege, Kaz, Daniel, Sebastian, McCool, Philipp, Cristiano, Dave, Farshid
Koster for the 1st part; Cristiano for the 2nd part
Logistics
McCool: no calls next week including the main call
… next main call will be on April 7
McCool: there are 2 weeks spring break coming up
McCool: planning to have the discovery meeting on Monday the 29th
scripting update
<kaz> Scripting slides
Spec status
Zoltan: summary of PRs and outstanding issues
… changes in InteractionInput and InteractionOutput data modeling
… uses either JSON value or readable stream
… JSON value if there is a schema, otherwise a stream with a content type
ConsumedThing and ExposedThing
Zoltan: the consumed thing changes include typing data input and output
Cristiano: should we check more conditions on validating consumed TDs?
… there is an issue with exposed thing as well
McCool: there might be 2 more stages of validation beyond schema validation
McCool: the categories are basic, intermediate, and advanced
sebastian: json schema is only one part of the validation process
… additional terms that are in JSON schema are allowed
McCool: we could use a stricter schema and require extensions to have a context file with a defined prefix
… also require assertions to drive validation
… there is JSON schema, then a script, then shape validation
Ege: the assertions are only checked for being present
McCool: we can annotate the implementation report with which assertions are checked
Zoltan: can we continue this discussion later?
… we need to know how this works for security
Zoltan: we need to discuss how are exposed things processed
… the exposed thing constructor is used to create a JSON object, then handlers are bound
… then expose() is called, then a TD is retrieved
… a big change was made, exposed thing doesn't inherit from consumed thing anymore
… exposed thing can be consumed using the consumed thing interface
… need review and input from the TD task force
… created an issue about how to create a TD for exposed thing
ConsumedThing
Cristiano: the consequence is that you need to use the consumed thing interface to invoke actions in an exposed thing
… thinking about bypassing the network interface when consuming things locally
Zoltan: for example using a local handler
Daniel: another reason for splitting this is that consumed thing properties are read-only
… and exposed thing properties may be updated
Cristiano: we don't have default handlers, now we need to set these handlers
Zoltan: it makes it harder to write identity functions
Zoltan: discussion about initialization of exposed thing
Cristiano: we use a dictionary as a partial TD and expand into a TD
… then validate and add required elements that are missing
… for example forms and security schemas
… cris: there should be a default form generation
Cristiano: there are issues how to handle title processing, whether it should be required
… partial TDs don't have mandatory elements
Sebastian: there is a question in the profile discussion about mandatory fields
McCool: title is more important than description because it's used in UIs
… also how do we handle language localization?
Discovery
Zoltan: discussion of discovery
… discovery is optional to implement, node-wot is planning to implement
… there is a common pattern emerging for discovery
… there is a request with filters and a wait for replies
… the API provides control to start and stop discovery and uses a cursor pattern to handle multiple outputs
… next step is API alignment and implement phase 1
… considering using the fetch API and readable stream data type
McCool: we should consider 2 discovery methods: automatic and direct
script management
Zoltan: runtime provisioning
node-wot update
Daniel: available releases are all marked v0.7
… working on a new branch with the new API
… there is a running instance for testing
… currently node-wot doesn't support discovery
… you need to provide a TD or location for the TD to be fetched from
… there is a management API for starting and stopping services
… management uses a TD
McCool: suggest setting up a meeting with security to discuss permissions
<kaz> (+1 to the joint discussion)
Daniel: node-wot has current support for many protocols and media types, some partially supported
… NPM statistics show a lot of interest and many downloads
Zoltan: there are peaks of activity around TPACs
Zoltan: any questions, discussion?
McCool: schedule the joint call for the week after break, week of the 12th
Zoltan: we could use the security call
McCool: short break and return for the plugfest discussion
<kaz> [5-min break]
plugfest
<kaz> Plugfest report slides
McCool: I'll go quickly through the results of the past pf. If you know more information please head to the issue tracker.
… we had a lot of different projects
McCool goes trough the list of the projects
W-ADE and A-MaGe
McCool: W-ADE and A-MaGe are from Ege
… W-ADE is an ID for TDs where you can interact with your TDs
… you can also expose virtual Things
… A-MaGe is abaout Mashup applications
… we tested those tools with coap thaks to Philip Blum
… we discovered that we can fetch tds using coap
… we might to add this to the discovery document
… semantic annotations are good to help the interaction
Testbench
McCool: then we have Testbench
… it's an evaluation tool to test online web things
… it was tested with different implementations
Ege: we pointed some possible problems in the slides
OneDM
McCool: tool for generating TM/TD from oneDM file
<sebastian> https://
McCool: we have alignment for using json pointers with SDF
… also there was a discussion about value scaling
… sdf has scaleMinimun etc.
… Michael Konster also experimented with the modbus protocol biding
Node-Red discovery
McCool: now node-red gen is able to automatically discover TDs in a TDDs
… and generate nodes to interact with it
ediTDor
McCool: now it supports TMs and some support for SDF
… the flatting process defined before in scripting can be used for TMs
McCool: ok that's it let's go in the next session
marketing
<kaz> [slides tbd@@@]
seb is presenting
Sebastian: we have a new webpage
… I really like it
… it's a good starting point
… where people can get in touch with us and with WoT topics
Sebastian: we also have an explainer video
… is on youtube
https://
Sebastian: we'll try to support multiple sub titles
… japanese, german, french, chinese, italian ...
… now we already have Chinese and Italian
… it would be best to have a double review for each language
Ege: it was an issue about synchronization in the original file
… is it solved?
… then also the newly added subtitle are not yet on youtube
Sebastian: yep we have to do it manually
Kaz: let's talk about the issue during the marketing call
Sebastian: we are glad to add other languages
McCool: the current supported list is good
Sebastian: maybe we should add spanish
agreement on publishing material
<Ege> wot-marketing issue 151 - Documenting the Twitter publication rules
Sebastian: there is some minor publishing material that it would be not feasible to get acceptance
… we also need to be reactive on twitter
… responses, retweet, likes etc.
… I'd propose a small subset of tweets that could skip the validation
Sebastian: what do you think?
McCool: how do we decide what's in the "boring" tweet set?
… probably we could authorize particular people to publish directly on twitter
… if others needs to tweet something they could refer to them
… chairs could occupy that role
Kaz: +1. There are several questions mostly about content and procedure
… boredom or excitement should not be really a deciding material
… I like the procedure described by mm. It would be better to define another term instead of boring :)
McCool: we should be careful for short text
… in meetings people can ask directly to chairs
Cristiano: what about retweet?
McCool: I think we should follow the same review process
Ege: I had a long list of questions about this
… could you please check it
McCool: let's document the process on this issue
McCool: email would be my preferred way
Ege: so this review process will apply for everything in the list?
McCool: yes
<citrullin> +1 on that. It's a lot of overhead.
Ege: my contribution with this process would be far less.
McCool: we could add Ege as a possible responsible
… this is something to discuss
Kaz: I know that I've been very strict, but we have to remind ourselves that we represent the whole W3C community.
Sebastian: don't we represent just WoT group?
Kaz: representing the W3C WoT WG means we represent the whole consortium
McCool: right
Dave: what about tweets in different languages?
Ege: I would not tweet in different lang
McCool: maybe we could have a dedicated account for each idiom
Daniel: I would not do that
Sebastian: also we don't have the man power to handle all these accounts
… plus twitter has the translate functionality
McCool: true, let's stick with english
Philipp: I understand the concerns
<Ege> thanks philipp! :)
Philipp: we should increase the tweet activity
McCool: we could give the authority to ege since is very active
Sebastian: let's sort this out today
McCool: ok first let's agree about the procedure
Sebastian: what about a warning system?
McCool: yeah, just ege be responsible please
Kaz: again please note I'm not saying we need censorship but just suggesting we should review the content before publishing it. we should be conscious about the CONTENT of the tweet... no misunderstanding, no spelling issues, no typos, etc.
<McCool> proposal: https://
<citrullin> I guess that could be considered to be a more critical topic
Lagally: tweets should contains a little bit of consensus. Although simply retweeting usually is not harmful as long it does not endorse a particular company or organization.
… if there anything that is controversial it is better ask.
McCool: is someone object to the current process?
Resolution: https://
Kaz: we should copy the procedure to marketing wiki
<Ege> https://
Sebastian: I had other topics to discuss like liaisons and how to increase community awareness
… let's discuss in the next calls
McCool: I'll send email with the cancellations about the next couple of weeks
Kaz: all the TF leaders, please also update the main wikipage
McCool: thank you for your participation of this virtual F2F
<kaz> seb: also would appreciate all your contributions!
<kaz> [adjourned]