User Interface Olympics Using Competition to Drive Innovation

Warren Robinett

Has bicycle racing improved bicycles? Yes, it has. We humans like to win, and like Lance Armstrong pedaling through the Alps in the Tour de France, we demand the best tools that can be made. The competition, the prestige of being the world champion, the passion to win, publicity for the chosen tools of the winners — these forces squeeze the imaginations of bicycle engineers and the bank accounts of bicycle manufacturers to produce a stream of innovations: lighter and higher-strength materials, more efficient gearing, easier and more reliable gear-shifting, aerodynamic improvements such as farings and encased wheels... the list goes on and on.

Competition spawns rapid improvements. Sounds a bit like evolution, doesn't it? Lack of competition can lead to long periods of quiescence, where nothing much changes. (Did you know the QWERTY keyboard was designed 100 years ago?)

This principle that competition spawns improvement could be applied to drive innovations in userinterface design. We call the proposed competition the User-Interface Olympics. Here is a sketch of how it might work:

• It would be an annual competition sponsored by a prestigious organization — let's say, the U.S. National Science Foundation.

• The winners would get prestige and possibly prize money (like the Nobel Prize, Pulitzer Prize, Emmies, Academy Awards, Oscars, and so on).

• The competition would be composed of a certain number of events, analogous to Olympic events. Individual contestants, or teams of contestants, compete for the championship in each event. Userinterface events would be such things as

- a timed competition to enter English text into a computer as fast as possible. (Surely someone can do better than the QWERTY keyboard!)

- a timed competition to select a specified series of items from lists. (Can we improve on the 40-year-old mouse?)

• Contestants would provide their own tools. This is analogous to the equipment used by athletes (special shoes, javelin, ice skates). However, for the User-Interface Olympics, the tools are the hardware and software used by each competitor.

• Since the goal is to stimulate innovation, contestants would have to fully disclose the working of their tools. A great new idea would get you one gold medal, not ten in a row. This is similar to the patent system, in which rewards during a limited period are bartered for disclosure and dissemination of ideas.

• An administrative authority would be needed, analogous in the Olympic Committee and its subordinate committees, to precisely define the rules for each event, for qualifying for events, and many other related matters. This Rules Committee would monitor the various events and make adjustments in the rules as needed.

• We would expect the rules of each event to co-evolve with the competitors and their tools. For example, the rule against goal tending in basketball was instituted in response to evolving player capabilities; in the 100-meter dash, precise rules for false starts must be continually monitored for effectiveness. Winning within the existing rules is not cheating, but some strategies that players may discover might not be really fair or might circumvent the intent of the competition. Of course, some competitors do cheat, and the rules must set reasonable penalties for each type of infraction. The Rules Committee would therefore have to evolve the rules of each event to keep the competition healthy.

• New events would be added from time to time.

These contests would be similar to multiplayer video games. The contestants would manipulate user-input devices such as the mouse, keyboard, joystick, and other input devices that might be invented. The usual classes of display devices (visual, aural, and haptic) would be available to the contestants, with innovations encouraged in this area, too. Most malleable, and therefore probably most fertile for spawning innovations, would be the software that defined the interaction techniques through which the contestant performed actions during the contest.

If we set things up right, perhaps we could tap some of the enormous energy that the youth of the nation currently pours into playing video games.

The rules for each contest, which would be published in advance, would be enforced by a computer program. Ideally, this referee program could handle all situations that come up in a contest; whether this actually worked, or whether a human referee would be needed, would have to determined in real contests. Making the referee completely automated would offer several advantages. Contests could be staged without hiring anyone. Computer referees would be, and would be perceived to be, unbiased. Early qualifying rounds could be held using the Internet, thus encouraging many contestants to participate. Figure B.14 shows a system diagram.

If this idea is to be attempted, it is critical to start with a well-chosen set of events. (Imagine that the Olympics had tried to start with synchronized swimming and sheep shearing!) A small, well-justified set of events might be best initially, just to keep it simple and try out the idea. One way to identify potential events for the UI Olympics is to look at input devices that currently are widely used:

• computer keyboard — suggests a text-entry event

• computer mouse — suggests an event based on selecting among alternatives

• joystick, car steering wheel — suggest one or more events about navigating through a 2-D or 3-D space provided by contestant provided by contestant

Figure B.14. System Diagram for a contest in the User-Interface Olympics, mediated by an automated referee program, with several contestants participating. The contestants provide their own hardware and software.

The real Olympics has events based both on raw power, speed, and stamina (weight lifting, races, and the marathon) and also events based on more complex skills (skiing, badminton, baseball). Similarly, the User-Interface Olympics could complement its events based on low-level skills (text entry, navigation) with some events requiring higher-level thinking. There are many kinds of "high-level thinking," of course. One class of well-developed intellectual contests is the mathematical competition. There are a number of well-known competitions or tests we can consider as examples: the MathCounts competitions run among middle schools and high schools; the Putnam Mathematical Competition run for undergraduates, and the math portion of the Scholastic Aptitude Test (or SAT, the college entrance test). Another similar competition is the annual student programming contest sponsored by the Association for Computing Machinery. One or more events based on solving well-defined categories of complex problems, using tools chosen by the contestant, would be desirable.

Strategy board games, such as chess and go, are another class of contests requiring complex skills. The rules for these games have already evolved to support interesting, healthy competitions and cultures. To focus on chess for a moment, by making chess an event in the User-Interface Olympics, we have an opportunity to reframe the false dichotomy between a human chess player and a chess-playing computer — we introduce a third possibility, a human contestant combined with her chess-analysis software. I personally believe that the combination of a good chess player, a good chess program, and a good user interface to integrate the two could probably beat both Deep Blue and Garry Kasparov. At any rate, this is a well-defined and testable hypothesis.

Therefore, the following events are proposed for the initial User-Interface Olympics:

• Text-entry speed competition

• Selection-among-alternatives race

• Navigation challenge: a race through a series of waypoints along a complex racecourse

• Timed math problems from the SAT (or equivalent problems)

• Timed chess matches

Each of these events would need precisely-formulated rules.

The strategy needed to achieve this vision of a thriving, well-known, self-perpetuating User-Interface Olympics that effectively drives innovation in user interface hardware and software is this:

• Fund the prizes for the first few years — let's say $100,000 for each of the four events

• Set up a governing committee and carefully choose its chairman and members. Give the committee itself an appropriate level of funding.

• Set an approximate date for the first User-Interface Olympics.

If the User-Interface Olympics were to become successful (meaning it had the participation of many contestants and user interface designers, it spawned good new ideas in user interface design, it had become prestigious, and it had become financially self-supporting), the benefits which could be expected might include

• rapid innovation in user-interface hardware and software

• recognition for inventors and engineers — on a par with scientists (Nobel Prize), writers (Pulitzer Prize), and actors (Academy Award)

• improved performance on the tasks chosen as events

Sometimes prizes can have an inordinately large effect in relation to the amount of money put up. Witness the prize for the first computer to beat the (human) world chess champion (Hsu 1998; Loviglio 1997). Witness the prize for the first human-powered flying machine (Brown et al. 2001). A million dollars or so in prize money to jump-start the User-Interface Olympics might be one of the best investments ever made.

0 0

Post a comment