Logging:
- Add basic syslog support (success or failure; number of updates)

Documentation:
- Write a man page
- Write a readme file

Mail:
- Use a different subject for success and failure mail
- Make receiving mail when nothing happens optional

General:
- Break out check-updates and download-updates into their own actions;
  move that logic to the cron scripts
- Re-do the lockfile bit so failures can be reported without resorting to
  relying on cron's output handler.
- Figure out what to do with systemd
- Update rpm, glibc, yum, python, specially?
- Check if we're running as root; exit nicely.