In message <[EMAIL PROTECTED]>,
    Ian Phillipps writes:

: On Thu, 25 Oct 2001 at 12:56:21 +0100, Robin Houston wrote:
:
: > I suspect that I have a strange idea of Fun (perhaps I should start a
: > list called "Fun with formal languages, complexity theory and
: > non-standard logic", where I'll presumably be the only subscriber).
: 
: Oh, I don't know  :-)

Me neither. :-)

: > That's what PSPACE-complete means. If a problem is PSPACE-complete,
: > that's a formal way of saying there's a snowflake's chance in hell
: > that anyone will ever find a general algorithm to solve it quickly.
:
: Is it related to NP-complete?

Yes, except it deals with space complexity rather than time complexity.

NP-complete problems are the subset of NP to which all problems in NP
are polynomial-time reducible (i.e., there exists a polynomial-time
transformation that can convert an instance of some problem in NP to
a problem in NP-compelte).  That means that a fast solution to any
problem in NP-complete would yield a fast solution for all problems in
NP.  (Note also that many encryption algorithms are trivially solved
in nondeterministic polynomial time.  Proving that P = NP would
destroy current protections offered by cryptography.)

PSPACE problems can be computed in a polynomial (in the size of the
input) amount of space, but the time is unbounded.  PSPACE-complete
problems are the subset of PSPACE to which all problems in PSPACE are
polynomial-time reducible.  A fast solution to a problem in this class
would be a miracle.

: > PS. There's a CSG for generating prime numbers (including a delightful
: > obfuscated C program) at http://members.fortunecity.com/brank/tor/primes/
: 
: .... but you can do that sort of thing with Perl's 'R'E, too:
:       grep {('x'x$_)!~/^(..+)\1+$/} 1..1000
: 
: or was that your point? The thread has passed me by :-)

Perl's regular expressions are very powerful and much more powerful than
what theoretical computer scientists mean by "regular expressions" (which
can only recognize regular languages).  For example, /^([ab]+)\1$/
accepts a particular CSL whose grammar would be a bitch to write.  (You
could use the pumping lemma for CFLs to prove that that aforementioned
language isn't context-free.)

Greg, having fun

Reply via email to