Knots or not?To summarize the issue: those in favor of using breaking knots argue that it increases the chance of holding a fall, those against argue that it takes time to untie the knots in an emergency. The discussion has been going on for years.
In Europe, it's more and more accepted these days to use breaking knots with a two men rope team on a typical glacier, and it's not a bad idea with three either. How climbers in other parts of the world see this, I don't know.
The more detailed argument is this: An unexpected and serious crevasse fall is hard to hold without knots. Even in practice, where, by nature, it's an expected fall (in a big crevasse), it's hard to hold because it's a big jolt. And the first and overriding priority if your partner falls into a crevasse is to hold him and not tumble right after him! Yes, it's cumbersome and it takes time to undo the knots to rescue your partner if they are in the way. But, bad as that is, it's really just a minor inconvenience compared to the alternative of landing at the bottom of a crevasse together. Even if the fall itself didn't cause serious injury, do you really think you can get out without help? I mean, on a typical glacier outing you'll probably only have two long axes and a couple of ice screws between the two of you. Good luck to get out of a serious crevasse with that.
Rope lengthIf you're a two man team with a 30 m rope, with, say, 15 m between the two of you and with three breaking knots, there is about 10 m of free rope left. Divided up, that's 5 m on each side, which is rather short to something useful with (depending on the technique you want to use). Much better would be a 50 m rope. The alternative to having free rope for a rescue is that the fallen climber prusiks back up again. Tiresome, but technically fairly easy - provided he isn't hurt in the fall, and knows what to do.
Only two climbers on a rope?On not all too steep glacier routes, the more people on the rope the better (up to a point), simply because there are more people to hold a potential fall. At least three is considered prudent. On steep routes, where falling isn't an option, the rules are different. Often two on a rope is better, sometimes you even unrope, for safety reasons. But from the questions I take it you're not planning to do that anytime soon.
DragAs for the rope drag on a the glacier? Never experienced that as a problem at all, regardless of the distance between my partner(s). I don't think it's ever been more than about 20 m though.
Below is an example of going unroped. The bergshrund on the right is visible, so that poses no serious risk. To our best knowledge the slope on the left isn't crevassed, but falling that way would definitely be very dangerous. However, had we been roped up and one of us had fallen, by the time the rope would be taut, the momentum would be far too much for the other to hold him. Therefore, our only chance would be if the one who fell would manage an immediate self-arrest. There are several ways to deal with this situation. We chose to go unroped - only advisable if you're
virtually certain that you won't fall, and are able to do an immediate self-arrest if you're wrong and fall anyway. If that fails, well, then only one climber dies, not two.