compression or limiting on synths that tend to go crazy on the loudness when cranking up their filters (tb303 or its emulations for example) is occasionally pretty nice to have in order to tame them in the mix. elsewhat they’re barely audible at low filter settings and drown everything else at high to no cut off.
besides that unprocessed drums + vocals are obligatory yelling for compression to gain their presence in the mix.
in the case of drums, the so called ny compression proved to be very helpful.
My mixing paradigm is shifting - I’m now using compressors on just about everything in favour of EQ. In fact, I’ve almost entirely ditched any use of peaking EQ. I’m finding a lot of ‘honest tone’ by using comps as unifiers. And if the sound is still annoying, well… ditch the sound. Don’t kill it even further with frustrated EQ curves.
If you haven’t tried it already try an open attack and a tight release - it’s great for a lot of channel applications. Use different compressors for different jobs too, they all behave slightly differently. That applies for both plugins and analogue gear.
Lots of great advice here!
Just to expand on my last post, you can see what I mean by use of compression in mixing by looking at what I did in this: http://m.thequietrevolution.net/mp3/mmd_ghost_debts_rehaunted.xrns - all native fx, 2.8.
Very impressive to make that track with only one hand drawn sample 
I think compressors can be very useful for mixing. But it’s importaint that they are used for the right purpose.
If you like a part in your mix to be more dynamic you can use a compressor to achieve this. To get more punch in a kickdrum you can use a compressor with 20-30 ms attack and a ratio 4:1. The attack setting means that the compressor only compress the sound after 20-30 ms and it gives the sound more punch.
I also use compressors on vocals and bass because this is parts of the mix I want the listeners to be focused on. So i compress them so they are more even in volume and sound louder in the mix. On vocals I use a ratio 4:1 and on bass I use a ratio 2:1. The other settings depends on the sound.
My experience is that you don’t need to compress everything and you must not compress the sound to much. I only use it on the main parts of the mix and then I want to fix a specified sound/part that nedd more or less dynamics.
My experience is also that sound from synthesizers don’t need much compression but if you record vocals, guitars or other instruments, they will definitely need compression to sound good in the mix.