Propose to use BG wiki, mailing list and repo
Ted: looking for task force lead[s], Glenn from Geotaba and Benjamin from BMW/Eurecom come to mind
Glenn: we are very interesting in being engaged and willing to help lead the effort
Benjamin: same
https://www.w3.org/community/autowebplatform/wiki/Data_tf -> potential areas of focus
Ted: I have left flattening out
the data model until the WG sends it over the fence
... signals might be available on different frequency than
requested in subscription or degrade under server load. we
might want to enhance VISS to communicate better the data
quality
Glenn: if it would be helpful at
the next meeting in two weeks, we would be happy to present a
telematics service provider perspective on data sampling and
quality
... we are collecting a couple billion data points/day and have
some specialists within that area
... I will work on that for the next meeting
ACLs
Ted: we have rudimentary security where what signals an application can access is controlled but that make no assumptions about on or off-board use. perhaps not in the spec but accompanying privacy guidelines that can be clarified, making it easier for an OEM to communicate to third party developers and enforce
Adam: the publishing model is
going to be different for the different use cases
... am I sharing a single value for display on the dash or
collection of signals for sampling
... is it time sensitive for v2x or more in aggregate for
diagnostics
Ted: how the data is collected, whether polling or using a known method such as Peucker should be captured and conveyed to cloud. you will want to know for example if speed is collected every minute or based on occurance of a significant delta change
Glenn: we would add sampling methodology to our presentation. we have things under IP we would be willing to open up
Ted: Benjamin, have you been looking at ACLs are part of your semweb work?
Benjamin: mostly focused on signals
Kevin: we are planning to implement details on access using security model in VISS spec and not looking at more at that
Transmission
Ted: I know some solution providers have used W3C EXI for getting data off of vehicles into the cloud, I will try to get more details. others in the group likely have experience as well
Glenn: we can potentially share our practices and not sure yet if they may be in or out of scope
Magnus: I would vote for having
it out of scope for this group
... we are seeing VISS being used on different protocols
already
Benjamin: I second that
... I know several are being tried at BMW and would not be
practical to tackle at this time
Ted: I will collect experiences to share and mark as out of scope for now
Business considerations
Ted: how aware of the marketplace do we need to be? various aspects of data collection could incur costs such as cpu usage and bandwidth and could be recorded and conveyed with the data payload. we should consider out of scope but something that might come back.
Ontology
Ted: at the WG f2f we agreed on flattening and evaluating the data model, cleaning up discrepencies Benjamin noted and furthering his ontology work
Benjamin: in a couple weeks I
will have more documentation available online and can present,
develop strategy on adding to data model
... on scope of where data model flattening should take place,
it crosses both groups in my opinion
... we want single availability, trust, etc and are all part of
a central data model
... we can align regularly with the main group
Privacy
Ted: Privacy is a rather significant
topic and we would likely work on guidelines more than
standards on the higher level. I will be involving W3C Privacy Interest Group and
regulators in the conversation.
... we also should be learning more about Caruso's consent model
that they presented at the F2F and are open to consider offering
for standardization.