in reply to Re^3: Infinite JAPHs?
in thread Infinite JAPHs?

This is true, but only under the assumption that only one JAPH can be composed per second. If we compose one per microsecond, then a far greater number of JAPHs is achievable.

One could, for example, write a program in - say - Perl, to generate a script with a random set of characters, execute, and check if the output matched "Just another Perl hacker[,\n]". It would then simply be a matter of adding processing power to increase your output (and assuming Moore's law still applies until the end of time, this would occur exponentially).

Replies are listed 'Best First'.
Re^5: Infinite JAPHs?
by dragonchild (Archbishop) on Mar 31, 2005 at 13:49 UTC
    Let's say we're talking 10**66 different JAPHs. The Universe is projected to live a total of roughly 30 billion years, or 3E10**10. That means the Universe will live (roughly!) 3E10**10 * 365 * 86400, or 10**18 seconds (with a little hand-waving in there to get a nice number). That means you would have to generate 10**48 JAPHS per second, starting at the Big Bang!, to get there. In case you're missing the point further, that's


    JAPHs per second. Or, the number of seconds the Universe is project to live . . . to the third power.

    Now, let's take Moore's Law. Let's say that it takes an average of a billion floating-point operations to test a random Perl script against /^Just another Perl hacker[,\n]$/. (We'll leave out, for a second, those that enter an infinite loop.) That means, currently, IBM's BlueGene/L computer (currently the fastest in the world at 92 teraflops) can do, roughly, 10**8 JAPHs per second. That's pretty good!

    Let's assume Moore's Law holds and computing power doubles every 18 months. (We'll leave aside the cost issues.) Assuming the Universe will live for another 15 billion years, that means that the computing power will multiplied by 2 ** (10**10). Now, it's a close approximation that 2**10 == 10**3. So:

    2**(10**10) ==> (2**10)**(10**9)
                ==> (10**3)**(10**9)
                ==> 10**(3*10**9)
                ~~> 10**(10**9.3)
    That means that the current 10**8/second will go up to roughly 10 ** (10**9.3 * 8) or, roughly, 10 ** (10 **10.2) / second - well above the projected 10**66 total JAPHs. So, yes, Moore's Law applied to BlueGene/L will handle the number of projected 4x80 JAPHs before the Universe dies. :-)

    Being right, does not endow the right to be rude; politeness costs nothing.
    Being unknowing, is not the same as being stupid.
    Expressing a contrary opinion, whether to the individual or the group, is more often a sign of deeper thought than of cantankerous belligerence.
    Do not mistake your goals as the only goals; your opinion as the only opinion; your confidence as correctness. Saying you know better is not the same as explaining you know better.

      Now all I have to do is convince my boss that investing in this project is a *good* idea!
Re^5: Infinite JAPHs?
by cog (Parson) on Mar 31, 2005 at 13:42 UTC
    to generate a script with a random set of characters, execute

    Generating a script with a random set of characters and executing it doesn't seem the brightest of ideas O:-)

    Besides, you'd be generating several times the same script (since the process would be random), and some possible japhs might never be generated (assuming a finite amount of time generating them).

      I wasn't trying to suggest the best solution, I was just trying to illustrate the fact that output of JAPHs could easily exceed one per second, even with a simple automated solution.

      But you're right, a more complex solution could be employed to increase the output even further (naturally at the cost of other things, such as demands on hardware).