triple_r wrote:
> Sounds familiar.  The job came down to a lot of signal processing, and I had a
> coworker who needed to improve his programming ability, so he decided to write
> a Sudoku solver on his own time.  Of course you need to keep track of which
> numbers are possible, so in each square for 1-9 a '1' represented a possible
> number while a '0' mean that possibility was ruled out.  Why not?  Makes sense.
>  But we need to store that data, so lets have '1' correspond to a 1, '10'
> correspond to a 2, '100' to 3, and so on up to 10^8.  Just add 'em up.  (so
> 1100 == 3 and 4 possible.)  Oh, and let's just store that as a real number
> since we'll be doing all those log10's anyway.  I was unable to convince him
> otherwise, which probably reflects more badly upon me than him, but at least no
> one else ever had to see it, and at least I think he used double precision...

Wait... are you saying that he implemented a bitmap, but in decimal, and 
stored it as a float?

Gah!!!

...Chambers

Post a reply to this message