- Include in the requirement differentiating processing time from copying out time.
- Make database table flexible enough so it can add new status if we want.
- Interesting files will have a taskID that way they will mark as interested ones.
- Include the batch jobID in the metadata.

- File status change: Do it when ifdh commands are called? That way users won't have to change their scripts.

- Make the thing zoomable.

Description doc on how the service interface should looks like?

Scope of the initial version:
- Record job and file states.
- Report state change to PMDB.
- FTS reports:
- Arrive to dropbox.
- Copied to pnfs and detected (declared)
- On tape (ready to be processed)
- Submit batch tasks.

Talk to Steve to check DB design.

Marc: we need to start the servers. Who is going to code what?

Traffic? We're expecting large traffic to need to be supported.