I record before and after.......I do a pretty light comp going in then once i've laid that track I do [ratio 2:0][att 1ms][rel 40ms] then im pretty much done with the comp

i do, mostly because i have a urei 1178 wich i love, and after my audio is in my daw, i prefer to stay all-digital..I like the sound of the urei compression..
But i must say that it's tricky...more than once, an unexpected loud passage has made my recording overcompressed, urei overcompression isnt very terrible sounding, but still the soundnbecomes a bit muffled and squased...pretty much irreversible..
so i'm more and more not really compressing while tracking, only gain reducing the very loudest parts..

Always...............depends on the singer how much. It seems to my ears that the analog compression going in helps make the digital compression (plug in) sound better on mix down.

As Billy Shakespeare hath stated:"Spank gently into the
good night"!!
Spanking the voice before A/D needs to be done, but not too
much...you can't "undo" analog compression. Try a 2:1, and then
run that into.....another 2:1!!! Two compressors in series seems to work better than one pumping a heavier ratio. Than you can have
fun in the digital realm! P.S.: You are probably using a condenser mic, right? You might discover that for certain "wild" singers, a
dynamic is better in that they tend to yield a bit of compression
on their own. Sometimes "bigger" ain't better.

- I prefer mic technique- don't use it at all (compression)when I record myself but on some singer you just have no choice-

I'm with the 'X' on this one. A decent sounding area for voice and quality mic technique gives me everything I need for a vocal pass.To say never, though, is inviting that one time.....One thing I do get out of a vocal track that has no artifacts is a clean dry signal and LOTS of headroom in my gain staging..If I was to use a compressor in the tracking stage it would be as an effect.In mix, however, there is no 'never'....only what is for that moment.

dynamics dyshmamics...

I guess that I am 'old school', but the talent that I deal with needs 'help'. I know how to read my meters while I'm tracking myself.But an excited singer...I want the PERFORMANCE...let 'em rip! I can spank it a bit and not worry about the emotion of the pass. I gave up a long time ago trying to get the vocalist to :
1) Sit still
2) Stop moving their heads all around while sustaining vowel sounds
3) Spitting sibiliants while spanking Susie on the seashore
4) Snapping their fingers, kicking the base of the mic stand,stomping their feet, slapping their butt ( that's MY job!), pounding their chests(well...sometimes...), and popping 'plosives on the plains...
4)

"Old School" method is to compress lightly when tracking to get max s/n and perhaps more at mix if needed to get the element to "sit right" in the mix. Different applications at different times.

Recorders that only had a dynamic range of 75 dB needed help to overcome sources that could have a dynamic range of over 150dB when closed miced. Even modern day digital systems can use a little help squeezing "real world" dynamics into a limited 125dB range. Soft knee at 2 to 1 taking at most, 6 dB off the top is a good place to start. I quite often only need to apply 2 or 3 dB of gain reduction to get the most signal possible to the track without overs. Less is more, be judicious and only compress as much as absolutely needed. You can always add more later if you need to to get it to sit in the mix correctly.

Coming from the "old school" myself, the old tricks die hard. I like to get the singer in close to the mic and work the proximity effect to fatten it up, to help with spill issues and to overcome the bad acoustics so often found in home studio tracking areas.

Even in a great room, I close mic singers and then if desired I will add a room mic six feet out from the first mic. A cool trick is to add more than one room mic, placing the second room mic another six feet further out from the first. Then run both the room mics through noise gates and set the threshold on each one to open up at different levels so that when the vocalist sings quietly only the close mic is recording. When they get a bit louder, the first room mic opens up and adds a bit of ambiance and a sense of distance. When the singer really belts it out, the second room mic opens up and adds even more ambiance and more distance ... this is an old David Bowie trick they used on "Young Americans" ... It works really cool sometimes in good sounding rooms.

I used to watch Al Kooper do the same thing with a Shure LeveLoc
and an E-V 666 in the room. What I was referring to was the fact that these days of American Idol-wannabe's, there are a lot of amateurs who have no idea as to how to work a mic properly. I try to be selective with what work I take on, but the truth is that $$$ talks...! And so, while I don't want to crush the hell out of the vocalist, I do have to rein them in at times...:)