Random but unique

miscdas at boxfrog.com miscdas at boxfrog.com
Sun Jan 12 15:46:01 EST 2003


[snip]
Dar Scott writes: 

>> On the final pass, the value of element 2 is swapped with the value of 
>> element 1.
> 
> This scrambling idea is a good idea and is especially applicable to 
> situations where the swap can be done in constant time as in scrambling 
> characters.  I think the random n of n can be generalized to random n of 
> m. 
> 
> I don't think the above quoted statement is exactly right, though.  It 
> looks to me that on the final pass element 1 is swapped with element 1.  
> On the penultimate pass, element 2 is swapped with either element 1 or 
> element 2.  Of course, the penultimate can be made the final as the final doesn't do anything.  This comment is intended only as clarification; the method looks great to me. 
> 
> Allowing R to be i (causing a trivial swap) is crucial and is part of the merit of the method. 
> 
> Dar Scott

My error regarding the final pass description. You are correct about the 
final and penultimate swaps as you described. Although the final swap is
a trivial swap of element 1 with itself, I left it to avoid another line of 
code. 

miscdas



More information about the use-livecode mailing list