I started off as a sysadmin, later becoming an Internet network engineer at an ISP. I got into Perl by finding 'Learn Perl in 21 days' or some such laying around the office, and read it (this was back in 2001). I then started looking at tasks that I had to repeat, and started writing small scripts to automate those tasks (my motto was that if I have to do something three times, I'm automating it).

Think in small terms at the beginning. Thinking great big is the wrong approach (imho), as you'll get ahead of yourself and get frustrated easily. Get the basics down extremely well, and you'll expand organically.

For instance. Say that every week you need to build a report based on a certain file, but first you need to extract the data out of the file. Here's a good opportunity to get a good foothold on some very basic (note I don't mean basic here as in it's not powerful, just that they're something every Perl dev needs to learn early) yet extremely important perl functionality: perldoc, open, split, strict, warnings, regex etc.

The more you practice on seemingly mundane and simple chores, you'll naturally progress onto more complex things. Every day, you'll run into something you don't know how to do, but that's what PerlMonks and other sites are for, after you've tried all you can and done some researching yourself.

Other good tasks at the beginning are copying/moving files, stopping/starting/restarting services, periodically checking memory/cpu usage, extracting/compiling tcpdump-type output etc.

Welcome to the world of Perl :)

-stevieb