> or to a point in time in a particular physical location

But how does this not take the local time zone into account? For "time at location", the local time zone is by necessity always involved in conversions, isn't it? There's just a difference between the time zone being explicit in your data representation, or merely implied.

But you cannot with any sort of confidence assume a future "implied" time zone, which makes turning a future timestamp into a local time, even using a timezone-naive representation, into an usure proposition.

Maybe I'm simply not aware of specific conventions around this topic, though, hence my original question.