As commented in #2, there's plenty of existing implementations that use the v4 timestamp, epoch and other parts of the data model. If we deviate too hard here, there's risk of impact into interop - but on the other side, new protocol allows for being able to do better designs, and the TAI/UTC requirements already change things up a little bit.
Consensus calls likely should be made on this subject within the working group.
As commented in #2, there's plenty of existing implementations that use the v4 timestamp, epoch and other parts of the data model. If we deviate too hard here, there's risk of impact into interop - but on the other side, new protocol allows for being able to do better designs, and the TAI/UTC requirements already change things up a little bit.
Consensus calls likely should be made on this subject within the working group.