Sorry for the late answer, I was mixing some Japanese band. I apologise for the language barrier stuff on my reply, because I am not sure of the word, but by "transient" I mean the first wave.
1) When trying to "smash" something (usually vocals/bass, as well as any parallel compression), wouldn't one always want to use the fastest attack possible? The way I see it is you're trying to even out the volume, so even letting a little transient through before clamping down seems like it would defeat the purpose.
If you just want it to sound smashed, then "Yes," but if you want it to sound good, then the answer would be "No". For example if you want to make the kick sound good and smashed, you will usually lose the punchiness of initial transient the kick if you put the attack to 0-1 ms, which is usually the lowest possible value. As a "soundcheck", ask the drummer to play constant 1/4 hits, put the release to 150ms, then start to twist the attack so that it sounds good. I think 5-30 is the one that usually sounds the best one, but once again this is really case dependant, use that one that sounds the best.
2) By the same token, when trying to "smash" something, why wouldn't one always want the fastest release? To me, that seems like it'd be the most transparent, cuz I'd imagine it'd be like manually automating the fader to adjust the volume of a part to even it out, which is what I assume using a compressor in this case is for.
No, because fast attack and release times on some compressors can cause distortion, especially on hardware units. Also really fast releases can make it sound really unnatural on high end listening enviroment, and for example the "default" setting of ~150 on most compressors is usually too slow on really fast double pedal action, because it can cause unwanted compressing of the initial transition of the next kick hit. Also if the release is faster than the attack, it can cause unwanted pumping.
3) When working with samples, especially those of a consistent velocity (specifically kick hits), what's the purpose of compressing them if you don't need to equalize the level? I mean, compression by definition is a dynamics control, so if the track you're applying it to has no dynamics, then what's the point? The only reason I can imagine is if you've got a compressor that has a certain kind of coloration that you like, so really, you're not even using it for its compression, just it's affect on the sound.
I almost left this one unanswered, but then I remembered when a DJ friend of mine said that when you compress it enough, it will distort. I also noticed this when we were tracking vocals at Finnvox, the distortion saturation was really nice with a high end compressor, it made the vocals sound a bit more raw. Also some compressors have a trait that they tend to make the body sound little fatter and more punchy if compressed correctly.
Also I found out on a tour that my cheap Behringer T1952 compressor sounds really fucking awesome on percussion type of instruments (read: drums), because of the additional colorization and slight distortion, which usually sounds good on drums. It beat the hell out of 8 times more expensive Klark teknik with same settings. But dont use it on anything else, it sounds horrific on vocals for example.