Why write a script to monitor a daemon?  Start it in /etc/ttys,
it'll get restarted if it dies.  Or use daemontools, I suppose.

Of course, finding out *why* it is terminating would be
preferred, ne?