Why would it be better to add gain rather than normalizing/compressing? In most mastering circles the former would be frowned upon, as far as I can assume.
Adding gain generally crushes peaks (clipping = bad). Normalizing to 0dB and then compressing adds overall volume without clipping.
Serious question!
Adding gain only crushes peaks if you add gain to bring your peaks up above 0dB. Normalizing to 0dB and then compressing will still leave you with levels below 0dB - how far will depend on the amount of compression you use.
At any rate, if you're compressing, you should compress first and then add gain/normalize to 0dB.
Uh, yeah...what Cooker said - he beat me to it. I will add, though, that I sometimes compress recordings with a very broad dynamic range because for most people's playback systems the narrower dynamic range sounds better. Of course, I always maintain the original copy and send the original WAVs to anyone who wants them (usually those with higher end playback systems).