Object identification

If the hierarchical structure of stationxml is flattened, there is likely a need for unique (globally or at least local to a document) ids for many objects.

For phase 1, for just network level, is the FDSN Sourceid (SID) sufficient, require the temporary network convention?

---

For later:

Items that probably will need ids:

• Network
• • just use FDSN SourceId?
• • enforce temporary network convention
• Station
• • Station epoch
• • epoch-less station?
• Channel
• • Channel epoch
• • epoch-less channel?
• Response
• • as part of a channel
• • independently, ex NRL
• SubResponse, like just sensor or just datalogger part, like NRL, that can be combined in sequence for a channel's full response

Perhaps a combination of SID and start time gives a good id for station and channel epochs, although I am not sure we can depend on the start time not being changing as mistakes are made and metadata can be updated long after data is recorded.

A syntax would need to be created, maybe something like:

FDSN:XD1994_GOMA@1994-06-02T00:00

MB: I am not 100% sure that fdsn:sourceids will do !

PC: Good point, both epochs and versioning may need to be part of any global ID. There is a long debate about using natural keys in databases, and this is a similar problem.