So there's already a Doom interface to killing UNIX processes:
http://www.cs.unm.edu/~dlchao/flake/doom/ .

What i'm thinking of is an abstracted set of library calls that will take
a complex "work problem," generate a corresponding "game problem" (like
the next block shape in Tetris), and use
the solution the player provides
to solve the "work problem." The crux that i haven't solved is this: are
there formulae that can produce a net savings of work by mapping the
work-actions to game-actions (and back), over just solving the original
work problem?

If not, maybe we could just collect entropy from Internet-connected gamers to improve rand().

ps. I did read _Ender's Game_. It reminded me of Netrek.

Sokobanhttp://www.cs.ualbe.../Sokoban/paper.html "Sokoban is PSPACE-complete" [johan, Mar 03 2000, last modified Oct 21 2004]

SOKOBANhttp://www.math.tau...dorit/sokoban.ps.gz SOKOBAN and other motion planing problems [johan, Mar 03 2000, last modified Oct 21 2004]

Minesweeperhttp://digitalmass....01/minesweeper.html The "minesweeper consistency problem" is NP-complete. [egnor, Mar 03 2000, last modified Oct 21 2004]

What do you think you're doing when you
play Solitaire on a Windows machine, merely
wasting time? Certainly not. A mapping has been made
between the intricacies of Klondike and various
well-understood software engineering practices.

Once the process has been bootstrapped, you need never employ
costly software engineers and testers again. The users generate the
next Rev.

There's some mathematical work on the
computational complexity of various
games. Some have been shown to be
P-, NP- and even PSPACE- complete.
(IIRC, Shokoban is PSPACE-complete.)
Not only _could_ we map any problem in
those complexity classes into those
games, at least in some cases the
mapping is already published.

That's a good point. I like my work (UNIX sysadmining) but i like playing games, and i know a lot of people who would rather be playing games than doing anything else. This would be a way all that crainiating that goes forth into computer gaming could be doing "real" work, and perhaps someone could make a living playing games, just like all those sports stars..