Meeting minutes
Agenda
Sebastian: scribe: Sebastian and Cristiano
<Ege> me, daniel, fady took minutes
Sebastian: Architeture,
Profile, (break), Scripting API, Marketing and Wrap-up
… any guest today?
Albrecht: Albrecht Schwarz
Sebastian: please be aware of
the W3C Patent Policy
… note that everything will be published
publicly
… so confidential information should not be shared
during this call
Albrecht: agree
Sebastian: next
… we're using IRC for taking notes
… also use q+ / q- to manage the speaker
queue
… resources
… wiki page here https://
… presentations available at
https://
… and then would like Lagally to take over the
moderator role
Architecture
Lagally: shows the agenda
for architecture and profile
… architecture is responsible for the abstract
architecture and interoperabiity profiles for WoT
… thanks to the contributors of the Arch and
Profile document
… Arch specifies the abstract architecture,
terminology and building blocks of WoT
… <shows toc of Arch document>
… <shows toc of Profile document>
<McCool> (it occurs to me that if these are just ed drafts, we need to have a FPWD by Jan 13!)
Lagally: lets go to the
Arch. document
… currently working on terminology
alignment
… some specifications have their own terminology
section, need to align and move common definitions to
architecture
<Ege> sorry for skipping the queue sebastian
no problem
<kaz> sk: we had similar discussion before
<kaz> ... we made the terminology section informative
I remember that PHL recommended to have terminology non-normative, however, I cannot remember what was the reason
<Zakim> kaz, you wanted to react to sebastian
Kaz: Terminology section have no nomative assertions. Personally I do not see the motivation to make it normative
<McCool> (I plan to mention the schedule constraint; we need to do all FPWD's by Nov 25 to hit CR by March 17)
<Zakim> kaz, you wanted to react to Ege
Ege: If the termonology section is normative, it does not prevent disalignment.
Lagally: thats right
<kaz> kaz: yeah, so from my viewpoint, what is important is not whether the terminology is normative or not, but the alignment itself
McCool: I do not have a hard opionen of that.
<McCool> (never mind, it seems both Arch and Profiles have had FPWDs already)
Lagally: another work is on
protocol bindings
… introduction section needs an update
… there also some places with editorial notes that
announce additional content
Lagally: next steps would be to work on implementation report, ...
Lagally: regarding
publication schedule it is planned to freeze the features Jan 31,
2022
… we expect to have the charter
extension
Lagally: thats about the architecture TF
kaz thanks
McCool: we should focus on Profile
Lagally: <gives an overview of the open issues in the Arch repo>
https://
Lagally: we need to take care on accessibility
its important when it comes to TAG review
McCool: we should also check internationalization
Lagally: I will create a issue that take care of it
<kaz> How to do Wide Review
Kaz: there is a guideline about the document review. We should follow it
Lagally: <creates an issue about group wide review>
WoT Profile
Lagally: what is a WoT
Profile?
… normative subset of WoT TD with a normative
binding
… <shows some drawings in the
slides>
just for clearification: profile do not develop specific binding like http, it simple reuse existing definitions
Lagally: <shows use cases
as drawing>
… important use case is the "out of the box
interoperability" use case
Ege: there is a problem with the "out of the box interoperability", even http is not interoperable since everyone can use it in own style
Kaz: W3C does not
provide a certification for interoperability, but we can and should
explain what is expected for WoT interoperability.
… we can provide a minimum set of requirements.
Also from my viewpoint, what kind of vocabulary to be used for what
kind of industry area is another important point for
interoperability.
Lagally: There are some core requirements such as for data model and protocol binding
there is not a 'common' data model
Lagally: right, I will change this to 'restricted'
Lagally: currently there is
work on the baseline for HTTP binding, identifying constraints and
rules on the data model
… inambiguous interaction semantics
… constraints on payload formats
… protocol binding semantics
… best practice security requirements
… compliance
<McCool> https://
<McCool> general
guidelines and considerations are here: https://
Lagally: there are open issues in the data model sections like about the data model constraints
<McCool> my suggestion is to lean heavily on the security-best-practices doc, which is currently being updated
<McCool> but even so we will have to define at least two contexts, because of the annoying situation with TLS on LANs
Lagally: regarding eventing model there is a 'Push' proposal
Lagally: Webhooks are used in many solutions like in github, paypal, atlassian and wordpress
Ege: we have no
imlementations of webhooks in the testfest so far
… we should also check what is implemented in
OpenHub as example
Lagally: there is the chance to have this in the Oracle binding
Daniel: I think, this is not the case. But Im not 100% sure
Kaz: I partly agree with Ege's points. We should ask WoT implementers about their approaches, e.g., ECHONET, NHK, the other Plugfest participants, Takenaka and Netzo.
Lagally: introduces Cloud
Events
… is a specification for describing event data in
common way
McCool: Im wondering if this implementable for the profile on time?
Lagally: for Oracle it would be important
Sebastian: want to ask about
the Cloud Events
… is that a standard or de facto?
Lagally: pretty small specification
<mlagally> https://
Philipp: it's run by the Linux Foundation
<McCool> (cloudevents is typical of "cloud" standards; someone publishes something, but is there an SDO? Not necessarily...)
Sebastian: any experience?
<McCool> (but if enough people use it... it's a "standard")
Lagally: my colleague is working
<Zakim> ege, you wanted to react to sebastian
Lagally: we have to work on the detail separately
<Zakim> kaz, you wanted to suggest we concentrate on the current features
Kaz: I'm wondering of the next step of the profile based on the provided MM schedule plans. I think it make sense to consider this for the next version. We can start initial survey in parallel, though.. A possibility is our going for Living Standards instead of RECs so that we can add any necessary features whenever we want but that would be a bigger question.
Lagally: we should plan this in the current release
Daniel: i think its
quite late for considering webhook and cloud events
… I like to understand what is the consequences
when we put those in?
McCool: we plan to have testfest Mid February
Philipp: we had a
simple discussion with SenML
… there was much discussion about this kind of
structure
<Zakim> Ege, you wanted to react to citrullin
Sebastian: our discussion on SenML was mainly about the payload structure
<McCool> (time check...)
Philipp: how to integrate it with TD
<Ege> ah sorry I remember now what I wanted to really ask
Lagally: the next steps are the push mechanism, rework data model, implementation report, compliance section, authentication and securiry, uncioorate PlufFest findings
Lagally: future work will
cover other protocols like MQTT, ECHONET Lite, OPC-UA
… provide requirements for TD 2.0
<Zakim> McCool, you wanted to react to Ege
Lagally: Publication schedule is to freeze Jan 31, 2022
<kaz> [10-min break; then Scripting API (35mins) and Marketing]
<Ege> cloudevents is from cloud native foundation instead of linux foundation right? @citrullin?
Scripting API
Daniel: new options for
consumer and exposer applications
… action state control
… discovery
… current node-wot status
Consume and Expose options
Daniel: pick one
particular security scheme or protocol
… currently there's an interaction level option
object to deal with particular binding.
… consume options will help with this
task
… is a non-breaking change
… similarly for the exposed side
… node-wot offer a lot of different protocol
bindings
… there's no practical way to configure the
exposing algorithm
… like specify what is going to be
exposed.
… but you need also a way to know is supported. In
other words, enumerate runtime capabilites
Action state and control
Daniel: there're some new
proposal about how actions should behave in the WoT world
… we are planning to introduce new APIs to control
Actions status
… we are thinking about a Control object to
manipulate an Action State
… it's a breaking change
… but we need the discussion to settle in other
spec (TD and Profile)
… there's no way to properly handle action
queue
… we need to understand the security model for
Action states
Discovery
Daniel: we are trying to
cover also the discovery spec
… we are waiting for a stable discovery
document
… we can alredy support basic discovery
spec
… however for scripts is beneficial to discover
things easier
McCool: regarding this we
don't have time to push new features in Discovery
… we can add this to the next charter
… it all nails down to which introduction mechanism
you want to support in Scripting API
Zoltan: the names are just examples
<Zakim> Zoltan, you wanted to react to McCool
Daniel: how to solve nearby?
McCool: bluetooth beacon introduction would work
Daniel: we are still considering how to split the two discovery phases
McCool: I would propose just second phase
Daniel: the issue here is that we still need introduction mechanism
McCool: we are missing
that
… we have to include on-boarding
Zoltan: I think the introduction belong to another API
<zkis> but we need to make a difference between discovering directories vs TDs
Daniel: I think onboarding could be specified inside Scripting API
McCool: let's have a joint meeting asap
consquences of recent changes
Daniel: ExposedThing does
not extends ConsumedThing
… how to properly test an ExposeThing?
… how to use it as a ConsumedThing?
node-wot status update
Daniel: now it supports
Readable Streams
… it took us a while
… it is going to be updated soon
… should we first implement than update
spec??
… what people think about this
approach?
… it would take longer to change the
spec
… contributions are always welcomed
… node-wot is increasing download count per
month
Lagally: do have statistics about users?
Daniel: we can't know for
sure
… but we are getting more
contributions
… we also have newcomers contributions
… also outside of the group
Lagally: it is good to see
Sebastian: we released in
march and April our new web page. Maybe that increased the
download.
… web page and youtube video
… if you look carefully the growth started right
after April
next goals
Daniel: allign with all
the TD features
… todo: canonicalization
… todo: discovery
… todo: wot profile? is it in scope?
… let us know if there's anything you would like to
see
Zoltan: we are focusing on interactions
Daniel: please join our call if anything
Kaz: I confirm echonet
used node-wot for plug-fest
… next point, even if Scripting API is a group
note, it would be nice to publish a new note soon. It would be nice
to publish updated Notes regularly when it becomes
stable.
McCool: scripting api should be consistent with node-wot implementation
McCool: we may not need canonicalization yet
Zoltan: we discussed about versioning and node-wot alligment
Zoltan: node-wot should be alligned with normative spec
Lagally: why are we withdrawing canonicalization?
McCool: we can't have it without breaking backward compatibility
Lagally: I'm raising serious concerns if the TD removes the canonical form. Profile is dependent on it
<Zakim> kaz, you wanted to suggest we discuss that during the Editors call
marketing
Sebastian: time issue... maybe it is better to split it
Ege: marketing task force
moderator changings
Ege: Sebastian resigned
as moderator
… after TPAC Fady Salama will chair
testing
… and Ege Korkan will take over
marketing
Kaz: I'm kind of surprised to see this "change" now. You should have mentioned this proposal beforehand. So would suggest this slide says "Chair Change Proposal" for today.
news
Ege: explainer
video
… improved web page
… new logo animation
… new task force descriptions
… sebastian contributed with a new documentation
page
… just a quick overview
… new tutorials by Philpp and a new one from
TUM/Siemens
… we organized all the videos in our web
page
… on twitter we are now 332 followers
… we have 6.1K impressions
… we now have 10.8K impression
… the video alone generated alone 13K
… content is critical to get good impression
count
Daniel: statistics are very useful, but we need information about website visits
<sebastian> +1
Ege: w3c is already
tracking but we don't have access to them
… no response from the w3c system team
Kaz: there were
actually responses from w3c system team
… we just did not reach the final agreement yet
sebastian wraps up the session
Daniel: there's an error in the agenda
Sebastian: true it is 3 hours
<kaz> [adjourned]