On 05/01/2008, Artem Baguinski <artm_at_v2.nl> wrote:
>
>
> The Rubik's cube interface as you've used it is a perfect example of
> languahge with very restrictive "syntax" - one is severely limited in
> what one can express in this syntax at the same time incorrect program
> is impossible to create.
That's interesting! With the Rubik's cube it's clear that not every sub-cube
can end up in any position but the same holds for those 4*4 sliding puzzles
with the one missing square.
Before plastic became cheap/popular/possible those used to consist of wooden
blocks in a wooden holder so one could cheat by removing all of the blocks
and re-arranging them. Somebody discovered that if you would swap two pieces
from a solved puzzle the result became unsolvable.
I googled this for convenience;
http://bd.thrijswijk.nl/15puzzle/15puzzen.htm
There are as many unsolvable states as there are solvable ones. We could
relate the unsolvable/impossible states to "incorrect programs".
What I'm aiming for, wouldn't it be fun to have something like Douglas's
Cube-system but aimed more at code then sequencing, yet with the possibility
to cheat included? Best (to my perverse mind) might be designing the system
to make cheating as tempting as possible. Cheating, unlike legal moves,
*could* lead to syntax errors which might propagate and eventually crash.
This would be quite hard to actually create but as a thought experiment it's
quite appealing and funny to me.
Yet funnier would be basing this on the Towers of Hanoi so one could use
that toy to solve a programming challenge instead of the other way 'round.
Yours,
Kas.
Received on Sat Jan 05 2008 - 14:17:27 GMT