D-Sane, we’ve been discussing two seperate phenomena (calles “problem 1” and “problem 2”) from the beginning of this thread (and the previous ones, since this topic just won’t ever cease to draw attention).
If you don’t want to read the whole thread, maybe start with my summary on the previous page.
While problem 1 (MIDI notes are printed at the time the key is pressed rather than at the time the (latency-delayed) sound is played during recording) is considered a particular behaviour rather than a bug (you would somehow expect the software to record the note exactly when the key is pressed), problem 2 (MIDI notes are printed EARLIER than when the key is pressed) is considered a bug.
You can check which one is your category by performing two simple tests (the “PLONK” test and the “record click” test). I think people should stop talking about problem 1 as a bug (and demand an optionally different behaviour instead, i.e. print the note when the delayed sound is played), while Steinberg should finally do something about problem 2 (which seems to happen most frequently with particular hardware configurations, which, on the other hand, tend to NOT cause the same problem in other DAWs).