<kaz> scribenick: Ege
<kaz> W3C Patent Policy
Sebastian: (showing the agenda)
... anything else for the agenda?
(none)
Sebastian: (review of the minutes)
<kaz> June-12
<inserted> scribenick: kaz
jun-12 minutes approved
Kaz: what about the vF2F minutes esp. the TD part?
Sebastian: would postpone to the next call or the main call because we're full today
Issue 97 - wot-binding-templates
Sebastian: work with IETF via
liaison
... Klaus will lead the discussion
<inserted> scribenick: Ege
Sebastian: The meetings about the RDF
Vocabulary of CoAP has been discussed
... the outcomes are documented in issue 97
Sebastian: we need to pick a new
slot
... for the TD call
<kaz> doodle
Sebastian: (shows the results)
... we want to pick wednesday 4-6 pm(CEST)
... it makes sense to have 2 hours
... is it ok for all?
... it is an improvement for the US east coast
... any objections for the new time slots?
<kaz> kaz: to be fair, we could start with this proposal and move the slot in 6 months if needed :)
<kaz> proposed slot on Wed
RESOLUTION: we will move the TD meeting to a new time slot. The new time slot is Wednesday 4:00pm-6:00pm CEST.
Sebastian: I am not there the whole next week but asked taki to chair the meeting
Kaz: next week meeting means wednesday, July 15 right?
Sebastian: yes, we start next week with the new slot
<kaz> @@@ slides tbd
Marcus: so I start
... here are some robots, that have different
capabilities
... ros allows to bind different protocols as a framework
... and to share code over different robotics
applications
... there are master, talker and listener nodes
... similar to MQTT but the nodes can talk with each other
without a middleman
Sebastian: what is the robot in this case
Marcus: nodes can be Thing or
Consumer
... so how can ROS be mapped to WoT
... invokeaction can be mapped to Publisher, subscribeevent and
observeproperty can be mapped to Subscriber in ROS
... so in slide 11, an example TD with ROS can be seen
... we see a URI scheme proposal, vocabulary proposal
Cristiano: what about defining contentType vocabulary like ros:contentType
Marcus: this is a nice input, I
didn't think of this yet
... I have a discussion slide on this. also there is a GH issue
as well.
<kaz> TD issue 912 on DataSchema and ContentType
Marcus: so adding ROS would be a nice addition to WoT since it is widespread
<Zakim> kaz, you wanted to ask if we can get the slides later and to ask about time synchronization among multiple "parts" and accuracy of the time
Kaz: Could you publish the slides on W3C WoT repository
<kaz> kaz: also it would be good to continue collaboration about time synchronization, time accuracy, geolocation, position of parts, etc., too
Cristiano: Did you work on a binding?
Marcus: so not really, I have written the TD and manually consumed and wrote the requests. I can work on it later
Sebastian: what implementation did you use
Marcus: Python with rospy
Sebastian: Could you make it public?
Marcus: sure
Sebastian: thank you for the presentation, it is nice to see that WoT can be used for ROS and be used to communicate with robots
<sebastian> https://github.com/w3c/wot-thing-description/issues/923
<inserted> scribenick: kaz
Ege: Philips Hue security scheme
[[
"href": "https://192.168.0.1/api/{{mytoken}}/lights/1"
]]
Ege: opened another issue since this is not official token handling
[[
"href": "http://example.org/thing/property/temp{?unit,api_key}"
]]
Sebastian: related to each other
... also need input from the Security TF
... is this notation related to the TD Template idea?
... or another mechanism?
Ege: IP address would be
templated
... Koster, any idea?
Koster: was using node-red for the security part
Ege: kind of a proxy
Koster: yes
... seems like HUE has some common security material
Ege: yeah, people who have
physical access can get apikey
... from my viewpoint, {{mytken}} is the only solution
... using '{{' and '}}' based on the template mechanism
... Twitter also does similar thing
Sebastian: we need security experts for
this topic
... McCool should be able to join the new slot on
Wednesday
... assigns issue 923 to McCool
<scribe> scribenick: Ege
Sebastian: Ben Francis provided the
differences between W3C WoT and Mozilla IoT
... issue nb. 921
... also he mentioned the WebThing community group
... We have remaining discussion
about defining TDT
... there is discussion on the architecture
... issue 454 on architecture
Cristiano: if we do this "TD is an
instance of one or more TDT" we need to define how it
happens
... like from a programming language
<inserted> scribenick: kaz
Kaz: as a starting point, this is good, but we need to clearly define "the same capability" at some point
<inserted> scribenick: Ege
Sebastian: why TDTs are needed
... mass production, digital twins, simulation, combining
different models
Ege: how would combining work?
Sebastian: it would happen in the development but you of course need to implement the backend
Koster: we need to be able to validate TDTs
Sebastian: analogies for TDT can be seen in programming languages
Cristiano: do you mean that TDs are not instances of TDTs?
Sebastian: sort of, so when you
instantiate the variable of a class, what you get is still not
a TD
... (shows extend mechanism)
Cristiano: TDs could implement multiple inheritance, which is not possible in many programming languages
<sebastian> https://github.com/w3c/wot-thing-description/issues/897
Sebastian: we need a way to use variables strings in TDs by defining where the values could be used, proposals show {{ }}
<sebastian> https://github.com/w3c/wot-thing-description/issues/898
Sebastian: we have a shacl shape that can be used for templating approach
<sebastian> https://github.com/w3c/wot-thing-description/issues/903
Koster: so you can go to examples in the SDF plaayground and these are mostly from OCF and similar other ones
Sebastian: So the next steps would be to
formalize TDTs
... and then SHACL to extend TDTs
Cristiano: I can try
Sebastian: we can ask victor as well
adjourned