Re: Anti-golfing

In article <[email protected]>,
	Peter Makholm <[email protected]> writes:
> In a momentary[0] insanity I came up with the concept of anti-golf.
> 
> Just looking for the largest solution wouldn't be fun. You can always
> add dead code. Demanding that every op-code should be executed isn't
> much better, then you just have to add ramdom not used calculations in
> the code.
> 
> For a while I thougt about the following rule:
> 
>   The program should break if you remove any subset of the characters.
> 
> This would solve both the above problems. But I guess you would still
> be able to make abitrary long code-snippet by some use of eval and
> some code/decode functionality. At least it would become arbitrary
> hard to prove or disprove that the program follows the rules.
> 
> Can anyone come up with a bounded measurement that can't be abused?
> 
Can't be done.

e.g. I can convert the program to a turing machine and run it on that.
then i can run that turing machine on a universal turing machine.
Repeat as often as you want.

If you use timelimits or memory limits, i just use other underlying
abstract machines that are faster/more compact

And there are no general programs to analyze general programs, so
whatever rules you specify, it will have cracks.