And, just a technique I often use: This may seem backward at first, but I've tried it both ways, and I like the results I get with this better:
Many people seem to like to put a limiter (or high-ratio compressor) last in the chain. I like to put it first. Well, not necessarily a super-high ratio, but a pretty high threshold, pretty fast response. Then, after that, a squishier opto-style compressor, with a soft knee and a low ratio, and a relatively low threshold.
My thinking is this: if you have your "gentle" compressor first, the high peaks are gonna hit it hard, and make it audible in a bad way, while your limiter will be limiting the already-compressed signal. And that doesn't sound great to me.
Oh, and moremagic, you're absolutely right that compressing in the electric guitar signal chain is an important part of many sounds; compressing before the amp has a great effect on the tone. I was thinking more about compressing the microphone signals used to record the guitar parts (coming more from an engineer's perspective than a guitarist's perspective, but your point is perfectly valid.)
Whereas, I prefer to let the high-threshhold fast-ish compressor catch just the loudest peaks, which feeds the squishy compressor a slightly tamed signal. (Coincidentally, this is a bit how tape reacts for me: the peak-to-average ratio gets smoothed out a bit as the signal is recorded, then your compressor has less work to do.)
"Your reality, sir, is lies and balderdash and I'm delighted to say that I have no grasp of it whatsoever." -Baron Munchausen