PatrickL: Benjamin brought two
topics for today, he is seeking experience with JSON-LD
... second is goals for upcoming TPAC meeting
Benjamin: WoT had a plugfest last
week, many impressive company efforts demonstrated
... goal was to create interoperability between these platforms
with WoT
... device has some metadata and knowledge about it, certain
capabilities (HVAC, engine, door in case of car)
... you can have interactions between them, perhaps events you
want to listen for
... they have schema for input/output
... they have certain amount of semantic information to be able
to identify what it is and its capabilities
... concept of heater (car HVAC) already known thing in
IoT
... you can include all the hidden information as part of the
meta data
... thing description for car is serialized in JSON-LD
[accompanying slides from WoT, will get URI from Benjamin after]
Benjamin: regarding representing
car we can use schema.org and VSSO representations
... when I parse this serialized data I get things like brand,
vin, etc
... I can get speed, action to activate ESC
... this serialization enables including these various data
types
... semantic types tell me about all the capabilities this
thing has, means to access (protocol and security
mechanisms)
... that is a short introduction to how cars can be represented
in WoT
PatrickL: it looks like a very capable way of describing what we are representing
Benjamin: normally would use
different serialization but those formats mean nothing for your
average developer. this is more intuitive for average
developer, familiar with JSON
... we should have a concrete look for our case
... we are already delivering payloads in JSON
Ulf: is this a replacement for VSS?
Benjamin: VSS could be used in
this case either as the construction scheme for URIs
... this is href system similar to your earlier naming
convention for elements
... VSS context would be available in thing description
context. it can be used to describe the thing and its
interactions
... ontology schemas are very powerful
Ulf: would this be stored within the car, in the cloud?
Benjamin: there are usually two
descriptions, a public more available to developers and private
with more limited interactions
... it provides documentation about what can be found at a
URI
... this serialization is pretty useful to provide semantic
annotations
PatrickL: regarding where it
would be stored, to me this is a description of the onotology.
it provides the scope of what the API could access
... I could see a W3C component and underlying company
solution. it would be possible to use the same mechanism to
expose additional (private) information
Benjamin: to expose for instance different regional capabilities for the same vehicle?
PatrickL: no, I could provide a
W3C view of the underlying vehicle information and a VW
specific one through the same mechanism
... how do you see this fitting our future needs?
Benjamin: it will likely become important in our data needs
Daniel: I am Daniel coming from
BMW, I joined the W3C WoT meeting and have been invited to
attend this as well
... we would like to see how this work is evolving, will join
you at TPAC and we are evaluating becoming involved
https://www.w3.org/auto/wg/wiki/F2F-201810
Benjamin: I went through the last
minutes and we had some high level ideas on groups we wanted to
meet
... most clear was Ted's email of potential groups we would
want to meet with
Ted explains liaisons with other groups, how we need to identify which topics, whether we would get a handful or warrants full cross group discussion
PatrickL: adding to the wiki
Benjamin was structuring, agree on JSON-LD, WebRTC
... regarding Devices and Sensors, do we want to meet with them
again?
Benjamin: depends who is available, they are meeting on Tuesday when we would be on data tf topics
PatrickL: what would be topic to discuss with them?
Ted: [gives recap on past interaction/influence] perhaps on adding new sensors, eg aftermarket, into vehicle
Ulf: for streaming data topic
PatrickL: adding to wiki
... you wrote TAG as well, do we have enough to talk to them
again? seems they are interested in a more concrete
description
Ted: I can try to get a few for Thursday on protocoless and v2 in general
PatrickL: WoT is obvious
... regarding authentication, I would really like to talk with
them about identifying users
Ted: agree, more a small consulting request or session on unconf day
PatrickL: we met with Web Commerce previously but seems like there is not much there at this time
Ted: agree, in short it is feasible
with some parts needing work such as transmitting securely
payments URI to vehicle
...a couple of the major payment providers
expressed willingness to explore if an OEM was willing. we can
revisit if there is later interest
PatrickL: re HTTPS in Local Network CG, do others see benefit in meeting with them?
Magnus: I think it would be useful, it is something we struggled with
PatrickL: agree we have similar
interests with WoT people
... anything for Linked Building Data CG?
... it would be interesting to hear from them
Ted: perhaps just keep an eye on them to avoid loosing too much time to liaison meetings
PatrickL: or send
delegation
... Ted will you reach out to them and place corresponding
blocks in our agenda?
Ted: yes, will do
PatrickL: ideas for next week's agenda?
Ulf: regarding TPAC agenda, didn't we say we should settle our major directions regarding version 2 and then start in earnest?
PatrickL: I remember that as well, it will be on the agenda
[adjourned]