sort items of myVar numeric \ by random(the number of items in myVar)
But should we count the number of lines or items first?
Here's a test to perform in a button:
on mouseup
--build a big list
repeat with i=1 to 10000
put i & cr after x
end repeat
--time the operation
put the long sec into t
--sort the original way
sort lines of x numeric by random(num of lines in x)
--record the time
put the long sec - t into a
put the long sec into t
--count first
put num of lines in x into n
--use the pre-count for the random
sort lines of x numeric by random(n)
put the long sec - t into b
--show the times
answer a,b
--show how much difference
answer a/b
end mouseupOn my computer the second sort is 28 times faster.
Still, the way it's written in the How To is easy to understand for introducing the concept. But for people seriously using it, here you go. I just think this is a beautiful example and I have done my own clumsy shuffling thing many times; you can bet I'm switching to this method!
(And actually, if you really wanted to be perfect, you might improve it by multiplying the number of lines or items by a desired amount for your random number parameter, so you have less chance of two lines having the same random number and (I assume) remaining in the order they originally were in the list.)
From time to time I'm going to see what else is lurking there in How To; I think it's not just for newcomers!!! :-)
Curry _______________________________________________ use-revolution mailing list [EMAIL PROTECTED] http://lists.runrev.com/mailman/listinfo/use-revolution
