It seems most people here are divided on whether a Torment sequel (spiritual or not) should adopt the new, 3e or later rules of D&D, or should go with a brand new rules set. So, I thought I would drop by and try do defend 2e and earlier a bit.
I mean, I understand the appeal of the later editions for a combat heavy game. A gold box like game would benefit from using 3e, or maybe even 4e, I think. The rules of these games seem to have in focus the "three pillars" of the D&D, mentioned in Mike Mearls's interview. So, the rules are more focused on describing how these three things, combat, exploration and interaction, should go. For example, 4e had rules to deal with traps by having the DM build them as an "encounter" (or something like that) where there is a series of challenges to be vanquished with skill rolls and other abilities the character possess. By the rules, this is certainly more interesting than simply rolling your Find/Disarm Traps percentage. There were similar improvements to the combat, by the rules.
But obviously, the point of combat and traps and what not in AD&D 2e, or 1e, or even old, plain D&D, isn't to simply play things by the rules. Yeah, simply rolling the percentage and determining if the players find or not the trap isn't much fun. And if you take the time to look at these percentages, most early level characters were pretty much screwed. But the point of those low chances wasn't that they were the default way of finding traps. They were just insurance, a little thing to help your character when something sneaked by you. Finding traps was more about paying attention to descriptions, examining unusual things and avoiding anything that might trigger it. Heck, sometimes you didn't even need to be a thief to do it. Similarly, combat and exploration had less well developed rules (specially old D&D) as you were supposed to come up with smart application of your skills, environment and weapons.
So, obviously, as
Brother None mentioned, a direct adaption of 2e isn't such a hot idea for a computer game. You would need to at least change stuff around. But what 2e does provide is lots of little hints as to how implement your game. Look at it like this, 3e and 4e focus on "tactical" thinking, stuff like "what are the best actions for me in the short term? What about the long term? How do I arrange my moves so I win this combat?". They are good, I think, systems (or at least starting points) if you want to make a game with this kind of gameplay. The older games, however, have some thing more like "adventure" gameplay. That is not to say you had a railroaded plot or puzzles that had to be solved in a specific way, like in many adventure games. But they were about creative problem solving, finding a way to use everything that is in the game to get what you want. They were all about Gordian Knot stuff, where to win you needed to think outside the box.
So, to foster this type of game, what the older editions provided were lots of hooks in the spells, items, skills, class descriptions or what not that could be fleshed out during play, both to challenge the characters and to provide them with ammo to face these challenges. While 1e took a much more well defined approach, it still had many hooks in there, but better yet, it showed many ways one could build on the hooks provided by the original game.The table about adjustments of weapons against ACs is an (often unused) example of this. The fire ball that grew until it filled its whole volume if constrained by the roof or the ground of the dungeon is another.
So, my point after this whole lot of text is that since 2e was more grounded in this "adventure like" gameplay (even if less than some of the previous editions), its hooks would be useful in the design of a more adventure like game as Torment was. Rather than having a tactical combat lie TOEE, for example, I think a simpler combat system, which allowed you instead to use different, unique actions on each combat, would fit the game better. Rather than putting traps that require different skills to disarm, using them as actual puzzles that can be interacted with different skills in different ways would be more in tune with that kind of game. And so on.