Re: compression algorithm is NP complete problem?

Indeed if you think about minimal space. That minimal usage problem is
known as knapsack. But if you think about the algorithm itself, and
the problem of logical decompressing, I don't know.

A practical aproach to NP complete problems might be thinking of a PC as
a deterministic machine (most certainly). So a BIG problem, such
optimization problems, may take exponential time solving the problem
(finding the optimal solution). An example is the salesman problem. If
you think about 10 cities, you can solve it in no time. Think about 100
cities, you'll die before the program stops (and your children will die,
too).

A practical aproach is knowing. I've faced some NP hard optimization
problems around there. E.g. a factory asks you to solve a problem that
requires optimization. If you identify the problem, and you say that the
problem is NP hard, then you can convert it into a NP c??mplete problem,
then find a heuristic and solve the problem. And you know that every NP
problem can be reduced to a NP complete problem. And if you look for
fast and efficient heuristics for those problems, you can solve your
problem in polynomial time, but with less accuracy.

I hope this is practical enough...
....and sorry for my bad english

.