Observation meta-data

[erikbosch]

In #313 we discuss meta-data on signal level. But would it also be beneficial in VSS to specify observation metadata?

As an example, VISS expects every value to have a "capture time", see https://raw.githack.com/w3c/automotive/gh-pages/spec/VISSv2_Core.html#read. Do we want to keep presence of "capture time" on transport level specification (e.g. VISS) or do we want to specify already in VSS that "capture time" is a recommended metadata.

One could think of other observation meta-data as well, like:

• Source (textual identifier on which physical sensor/system that provided the value)
• Accuracy (Absolute or Relative for given value)

Theoretically signal-specific or branch-specific meta-data could exist as well, concerning e.g. which method/algorithm that has been used to calculate a specific observation. As an example, there might be multiple providers

Is this something that would make sense to specify in VSS, or is this something that possibly better fits in VSSo. (In VSSo they already have propertyValueUpdatedAt, see https://github.com/w3c/vsso/blob/gh-pages/spec/vsso-core-re.html#L311)

[erikbosch]

Meeting notes:
Sebastian - we could define a set of fields/labels that we recommend to use. But needs to be discussed if so how it shall be marketed.
Erik - maybe we want to propose recommended fields/profiles for layers, for e.g. authentication
Ulf - Not sure that it shall be in VSS. In VISS we have metadata keynames, if VSS wants to adopt them it would fit VISS, but other APIs may want to use different.