Ahem. I would also like to point out two things:

First of all, there's numerous code instances like this:

if ( pSoldier->bTeam == ENEMY_TEAM )
{
	switch( gGameOptions.ubDifficultyLevel )
	{
	case DIF_LEVEL_EASY:
		sBreathPerAP *= 9/10; // -10%
		break;
	case DIF_LEVEL_MEDIUM:
		sBreathPerAP *= 1; // normal
		break;
	case DIF_LEVEL_HARD:
		sBreathPerAP *= 11/10; // +10%
		break;
	case DIF_LEVEL_INSANE:
		sBreathPerAP *= 6/5; // +20%
		break;
	default:
		sBreathPerAP *= 9/10; // -10%
		break;
	}
}
// SANDRO: this simple thing is the correct calculation
sBreathChange = (sUnusedAPs * sBreathPerAP);

In these instances, the difficulty level is used to determine some modifier. If none of the original levels is used, the game will now use the novice settings.
This makes the entire idea of externalising difficulty levels settings useless. If you define a new difficulty setting and copy its settings over from the INSANE tags, you will get a weird hybrid of INSANE and NOVICE settings.
This isn't strictly speaking a new error - it just wasn't relevant until now. If someone wanted to play with a modified INSANE or NOVICE, he'd likely start with that anyway.
This can be fixed, but this is a design flaw, not an error. Fixing errors is ok as everybody does them, but I don't see why coders should finish features of other people.

Second, several LUA scripts break if one starts the game with number of difficulty levels = 4. This can be fixed, but leads me to seriously doubt whether this was tested.

Also, I ask again: Will there ever be any explanation of this thing? At least a post announcing it to the public?