This is a two-part question.

1. The duration of some spells (cleric kit spells come to mind) are expressed in seconds per level (six seconds for cleric kits), whereas others are based on rounds per level. Given that a round is six seconds, is there any difference?
2. I read somewhere (and I don’t know if it’s true) that improved haste works by cutting the time of a round in half, so that two rounds are “squeezed” into one, thereby effectively doubling the apparent number of attacks per round. If this is true, how does this affect those spells where the duration equals a certain number of rounds – i.e., is the spell duration basically cut in half?

1. No difference.
2. It does, but haste effects don't affect spellcasting--duration and casting times won't change.

Thanks, CamDawg. I can always count on you! :)

Also, thanks for the confirmation of how improved haste works - i.e. squeezing two rounds into one. I guess that also means that, when dual-wielding with IH, you'll get two off-hand attacks in a round.

Thanks, CamDawg. I can always count on you! :)

Also, thanks for the confirmation of how improved haste works - i.e. squeezing two rounds into one. I guess that also means that, when dual-wielding with IH, you'll get two off-hand attacks in a round.
Haste twist #2 is that it kills all offhand attacks. Nythrun probably has the best example of the inner workings of haste in her Fixpack research (http://forums.gibberlings3.net/index.php?s=&showtopic=7531&view=findpost&p=65112). (Haste and Improved Haste internally use the same opcode, so the post applies to all haste effects.)

edit: BioWare was lazy and used the haste opcode in many places where it simply wanted to increase ApR or even movement speed (as opposed to actual haste) and this caused no end of unintended consequences and odd bugs. BG2 Fixpack--like Baldurdash before it--swaps many of these with other opcodes to achieve the same ends. For example, the Boots of Speed are altered by both packs from a haste opcode to a straight movement increase bonus, in line with their description.

Is it just me having problems, or is G3 down?

Anyway:
Haste twist #2 is that it kills all offhand attacks
That's odd, because I seem to recall a thread where someone had tested this and they had two off-hand APR's. Then again, I could be mistaken (which is why I asked the question in the first place). Are the lost off-hand attacks replaced by main-hand attacks?

I've also renamed this thread because it's expanded beyond the original question.

Is it just me having problems, or is G3 down?
Nope.

That's odd, because I seem to recall a thread where someone had tested this and they had two off-hand APR's. Then again, I could be mistaken (which is why I asked the question in the first place). Are the lost off-hand attacks replaced by main-hand attacks?
That'd be the first I've heard--every test I've seen shows that the offhand attacks are killed. Keep in mind that this is with the actual haste opcode; there are several ways to make haste-like effects in the engine without an actual haste opcode. With a 'faked' haste the offhand attacks should be perserved.

As for whether they're replaced, check out Nythrun's post I linked--you'll get (at minimum) double the main hand attacks.