> > Bernie Cosell wrote:
> > My intuition is that the problem is too constrained
> > and that there  *is*no* algorithm that meets all the
> > criteria, but I'd be delighted to be proven wrong...
>
> Simon Cozens wrote:
> The original problem - finding a random line in a file
> - was solved by induction. I tried to apply something
> similar to this problem. n=1 is easy. What does it mean
> to take the middle line of a file when n=2?

I honestly think I've found the only true solution, which
is:

perl -p

I.e. the only storage you require is for the current
line... which is just printed.  Meets requirements AFAIK so
it looks like a trick problem!

Jonathan Paton

__________________________________________________
Do You Yahoo!?
Everything you'll ever need on one web page from News and Sport to Email and Music 
Charts
http://uk.my.yahoo.com

Reply via email to