It works 50/50. Its great when you’re standing on the ground and you want to mantle up somewhere; its not too great when you want to latch onto a ledge in mid-jump. However, Thief 3 handled that very well so it can be done right with the proper world structure.
That is for 3D games, 2D games should copy how Prince of Persia did it. Perfection (IMO).
Thief mantling is a perfect example of the developers understanding that a game is not a wrestling match with the controls. The controls are there to you help you achieve what the game will let you do; not get in the way. As many games nowadays attempt to tend towards realism it is a mistake to try and expound upon the control system to let you attempt to control every detail of what your avatar can act upon in the game world. In “Real Life” we do these things not without thinking but because we have had years and years and years of being immersed in an environment which requires our detailed interaction. You can’t cram this into a game. Apart from your game world being hopelessly inadequate physically, a mouse and 100-odd keys is utterly incapable of replicating the hundreds of analogue-driven muscles you’ve learned to control with a thing so powerful we cannot even hope to replicate its function within our own lifetimes - that’s your brain.
Games which are designed around mastery of controls are by and large of extremely limited appeal to a small minority of people, in the same way that being really good at ping-pong is of extremely limited appeal to most people. Platform games and shoot-em-ups are massively unpopular compared to their heyday now because of their total focus on reflexive control with reward largely and in many cases solely coming from mastery. This trick doesn’t work any more.
So: instead we attempt to capture the intent of what the player wants to do, and then work out the best and most intuitive way to map this to the minimum possible number of controls, and even then, we do all sorts of little hidden things to make the player succeed in what they were trying to do if they’re not quite got it right.
Cas 
Mass Effect 3 got on the “mash butan” mechanic too: sprint, dive, cover, and mantle are all done with the spacebar. The last two cause a bit of a problem if you mis-time it; I’d accidentally vault over cover instead of ducking behind it, and end up staring the enemy right in the face. Oh well, that’s what the omni-blade is for. “Hi, name’s Shepard, nice to meet you” tinkletinklezrrrch glitch
That would be the disadvantage yes - too much of the same thing leads to “conflicts of interest”.
Still I’m an advocate of being able to do all your repeating actions with at most two “action” buttons. I like how Limbo does it; only one “do something” button and the game never ever has to tell you when to use it. The game in stead invites you to try and do stuff by making you curious. There is something sticking out there that looks like I can pull or flip, let me try what happens… Hmm, the area is pretty much rounded but there I see a sharp corner, I can probably climb up there if I jump to it. Etc. etc. The controls there work because the world around you makes a whole lot of sense in stead of the other way around: the world starting to make sense because you understand your control capabilities.
To be honest one button control is best, but a lot of really nice little game designs spring forth from having two. I suppose this is why it’s universally accepted that a two-button mouse is the norm. Oh wait, unless you’re on a Mac…
Cas
True. Two, not three or more. As soon as games start to bind functions to a third button, I start looking for keys on the keyboard. On the other hand games that combine button presses (IE. hold right mouse button, click left one) don’t bother me at all for some reason. As long as it sticks to two buttons.
I’m revising my statement: max two action buttons… per hand. Movement also counts as that is generally done with no more than two button presses at the same time (to move diagonally, to run or to sneak as examples of two button actions).
But I hate the games where there’s a different key for everything. Using the same keys/buttons might be more inconvenient, but it’s a lot easier to remember and follow (like Doom).
Weird example. Not only was Doom itself far simpler than the average action game (no jumping, no sneaking, no looking up or down), it has a dedicated button for everything; shoot, run, strafe, activate, map, …
Doom is perfection in FPS control for beginners. Quake consolidated the controls for all future FPS games to come.
Cas
I wish devs would relearn it again though - Quake was indeed near perfect. Don’t need an action button to push buttons, just bump into them or shoot them. Brilliant!
Of course ID software learned it from Nintendo while doing Commander Keen, so we should all look at Mario to be honest.
I recall that Carmack argued strenuously against the concept of a “Use” button after Quake but was overruled. Which is a shame, because IMHO he was right.
Andy Schatz did a very interesting article about the design of Monaco recently and how his experiences with Apple made him design the game - read it on Penny Arcade here.
Actually just reading that article through again basically just summarised my whole POV on the subject I posted earlier
Cas
I quite liked the seamless gui stuff that Doom 3 did, it allowed interesting in-world interfaces without being jarring or needing a use key. You can see them used to enter a combination at the beginning of this vid:
Anyone know any other games that did anything similar?
All FPS games suck with a joystick/controller. The only good way to play them is with mouse and keyboard.
The best controller for FPS is mouse+keyboard. The best game for Xbox controller is FPS. Playing PES with xbox’con isn’t funny. They made it with two triggers, so you can shoot with dual gun and go play Halo 2 or Darkwatch!
It really depends on the type of game, there really isn’t a single input device that fits all needs.
I’d say that is true for fast paced FPS games (Quake, Unreal Tournament, etc), however slower FPS games (Halo, Golden Eye, etc) work reasonably well on a game controller.
RTS games (Starcraft, Warcraft, C&C), especially those requiring high APM really do work best with keyboard and mouse, don’t see another controller scheme beating them any time soon.
Beat em ups (Street Fighter, Soul Calibur, Virtua Fighter, King of Fighters, etc) work best and most naturally with an arcade stick, there really is no going back to other controllers once you’ve mastered the arcade stick. Playing these type of games with a Game pad will pretty much destroy your thumbs.
Game pads are typically a slower input medium than the above specialist controllers, however they do give much clearer, simpler and slower feedback to the player from button press to reaction on screen (the original nes controller design is a nice example). Its also why they are mostly used in game systems aimed at the masses (e.g. consoles, handhelds, etc). Some people just require input to be slower as its less intimidating, easier for them to pick up and understand. Its also why I think most games these days have in game buttons/switches that you have to manually and consciously press to trigger. Personally I do prefer the more streamlined approach of quake style buttons that trigger automatically on touch or on being shot but I can see why it can cause confusion for certain type of players.
I have actually literally burned my fingers playing Beat em ups with pads. Of course thumbs are the first to go. Just do 30 Hadoukens in Street Fighter and you can already feel it.
I have only one response to that.
“Cheesyyyyyyy!!!”
I can’t even aim fast enough in Halo with the gamepads. You have to move the joystick just a touch. Oops overshot. Move back. Damn overshot again. Ad infinitum. Or worried about overshooting. Inch a bit. A bit more. A bit more. A bit…dead.
True enough. Gamepads and joysticks are pretty much equally bad, but in FPS’s joysticks have the advantage.
I started playing LOTRO last night and the WASD controls feel very grainy. Trying to get in front of a spider to whack it felt less than heroic and more like parallel parking. I think the spider died of laughter.
Also, the camera tends to get stuck in a tree while I am busy parking in front of a very patient monster - I wonder if there is a way of not drawing scenery intervening between the camera and the avatar? It can’t be easy because it seems all these games have the same problem. Maybe a pixel shader in the middle of the screen that lets through the image of the avatar? I don’t know what I’m talking about, yet 