There is one design aspect above all others that can sink an otherwise excellent videogame. The graphics can be astonishing, the architecture beautiful, the AI fearsome - but if the player is forced to negotiate a bad control system, all that is next to worthless.

Control systems - or HCIs (human-computer interfaces) - can be bad in two major ways. They can be simply clunky, or actually illogical. Shenmue’s is both at once. Character movement is assigned to the D-pad in the “rotate/walk forward” paradigm familiar from Tomb Raider, Resident Evil et al. Now this is an elderly and clunky system that the N64’s analogue stick was supposed to consign to history. Four years on, to relegate the Dreamcast’s analogue stick to the secondary function of “look” is nuts.

But Shenmue’s control system is actually illogical as well. There is a good reason for the rotate/walk system in the early Tomb Raider games, which is that success depends critically on very precise positioning and jumping. Later sequels introduced analogue movement with the Dual Shock pad, but the D-pad was still regularly needed. Yet no such acrobatic negotiation of 3D spaces is required in Shenmue, so there’s simply no excuse.

Another way in which control systems can be illogical is in a structural mismatch of physical action to virtual action. In the Dreamcast port of Quake 3 Arena, it makes sense to assign looking and turning to the analogue stick, because manipulating a joystick is a good structural analogy to moving your head around in the virtual space. But you’re then required to use buttons to run and strafe, which is a bad method of controlling human movement. In real life, we do not always run in the direction our heads are pointing; nor do we restrict our movements to two axes on a plane according to the direction we are facing in.

Now you’ll tell me that it doesn’t really matter, as anyone who is serious about playing Q3A on Dreamcast will buy a keyboard and mouse. To me, that’s a fabulously regressive argument. Keyboard/mouse control only came about in the first place owing to the dearth of standardized PC peripherals: it is a method of (mis)using hardware that was never designed to act as a game-input device. Of course it is currently the most efficient HCI available for Q3A, Unreal Tournament et al - yet that’s not natural cybernetic justice. It’s not because it is a good interface design, but because these videogames were coded from the ground up with that flawed HCI in mind.

Yes, I am arguing that keyboard and mouse is aesthetically inferior to joystick control for FPS games. Partly because, as with Dreamcast Q3A, simple binary button-pressing is an inappropriate system for movement, and also because of another, more serious, structural incoherency. Trying to aim a gun in real life is not a matter of “pointing”; it requires you to direct the gun and use small corrective movements to ensure that the barrel is aligned appropriately along three axes. Such directional control is much better approximated by the use of a joystick than with a mouse. The mouse-controlled reticule privileges no part of the space over any other and requires no more effort to target an enemy half a screen away than one who is a mere half-degree off the centre of your visual field.

So mouse aiming is far easier than aiming with a stick. And to compensate, PC-based FPS games are made artificially more difficult in other ways. Of course, the maniacally swift movement, 1080-degree split-second turns and multiple frags of such games are highly entertaining in themselves, but in terms of suspense and player satisfaction, the console-based HCIs - especially the sweetly engineered dual-analogue control of TimeSplitters - remain superior. When aiming is more difficult, hitting the target becomes that much more satisfying and enjoyable.

To add another example, if Dreamcast Silent Scope is played with the mouse, it becomes stupidly easy compared to the directional joypad control. But further, joypad control was an utterly bizarre creative decision in the first place, when there exists a first-party Sega lightgun. Some system of aiming with the gun and using a button to call up the sniper scope could easily have been implemented, and one can only assume that it was political considerations - fear of the hysterical anti-videogame lobby in the US -that prevented it.

The corollary to these arguments about poor interfaces is the fact that new sorts of HCI open up new gameplay possibilities. That is why Shigeru Miyamoto decided that Mario 64 required a (literally) revolutionary new type of controller. But the danger now is that HCI development might be stagnating. Sony copied Nintendo’s analogue stick, but at least they added another one. Creative twin-analogue systems have only been tried in a very few games - the terribly frustrating RC Stunt Copter on the one hand; the brilliantly inventive and underrated Ape Escape on the other. But it is disappointing that Nintendo and Microsoft have simply followed suit with Dual Shock clones for their new systems.

One ray of hope was provided at last year’s ECTS, when Sony had a concept stand demonstrating the use of giant foam peripherals, such as swords, whose movements were interpreted by a simple USB webcam and translated into the movements of a virtual sword on screen. This is one way in which good HCIs can also open up new markets. The clunky D-pad control of Shenmue is an instant barrier to the game’s artistic merits for someone who hasn’t spent years fiddling with similar control systems. But Samba de Amigo is a great hit at parties, with women as well as men, non-gamers as well as gamers, because the interface is so wonderfully, instantaneously attractive.

From the clunky and illogical results we tolerate, it seems that a game’s control method is often only finalized in the relatively late stages of development. But to be truly great - and this is as true of Asteroids or Robotron as it is of Samba de Amigo or TimeSplitters - videogames need to be built around a well-conceived interface from the start.

Leave a comment: