I am creating a news crawler script that will update my website with relevant news etc. Im not sure how to run my script though, i want it to run 24/7 and not rely on me being at my computer. I was wondering if i should use a cron or execute as a shell command so it runs forever.

The cron option is good except i'll have no idea on how much it'll download and how long it'll take, plus if it needs to follow links it might go on for a bit.

The shell command would be ideal, but im on a share server and i'll probably get a telling off for running the script 24/7 and consuming the server's juice.

Any ideas?

Cron is a good idea.

I have scripts that run every 1 minute.
However if that script takes longer than normal i have it die.

I normally put a LOCKFILE in its directory..

Ie:
when it starts, It creates a file with a value in it.

When it finishes it removes the value.

When the next one runs, It checks for the value... if its there, it dies.

I then have another script that checks the creation time of the lockfile.
If its been locked for more than 30 minutes. I get a mail so i can check why its still there.

Thanks, that's just given me a few ideas.