... Frankly speaking i am also searching for the answer!
I had just started reading some of the initial pages of New Kind of Science by Stephen Wolfram
... and i have got a feeling that well it might be possible to write any software in a "more simple" way but not simpler.
For example i was trying to write a shortest program (in terms of condition checking) for finding all possible knight moves in fortran and ended up with the following, o/w a lot of thinking:
data id /2,-1, 2,1, -1,2, 1,2, -2,-1, -2,1, -1,-2, 1,-2/
read(*,*) i, j
thought i feel the code is compact, i have a strong feeling that the program can be reduced further. but to do so i have to do a lot of background thinking, which infact may complicate the understanding of the program it self!
... may be my conclusions are primitive, or completely wrong? ... i am still researching!