Shuffle playlist but keep order?

Question

I'm using iTunes to deliver stimuli for an experiment. There are 36 aural examples that subjects will listen to, and I've got them in a playlist in iTunes.
I want iTunes to randomly shuffle them. I know how to do this. That's not the problem.
There needs to be a pause between each track -- that is, the playlist can't automatically go on to the next track; the experimenter needs to click on the next audio example to begin it. This is also not a problem. I know how to do this too.
The problem is that when iTunes generates a random order for the stimuli and plays the first example, and then stops, when I click on the second example in the playlist iTunes reshuffles the entire playlist.
How can I get iTunes to shuffle the playlist once but not re-shuffle it when I click on the next audio example?
I'm running iTunes 5 on OS 10.4.11. I know that's an old version, but it's an old computer I dug up to run the experiment with and I don't have the resources to update the OS right now.

Answer 1

I guess what I meant was I was looking for the way to
do this that would result in the highest performance
as this is a common recurring query.....
What I was initially thinking of doing was maintaining
2 Lists - one of which would be a shortened version
with just the 'compound index fields' and it's index
position in the Full List. I would then propose
sorting the short list once all data has been read,
then scanning through it for 'duplicates' and only
updating the Full List with any necessay changes. This is awfully complex. And if, as you say, the duplicates are infrequent, then it's not going to be very efficient. It's generally best to touch data as few times as possible, and since you're touching it once when you read it, that's a good time to do the dup check.
I'd prefer not to define an object to be an
abstraction of each row since only the index
columns(mainly) remain constant. I'm assuming that you're going to actually use this data in your program, right? In that case why wouldn't you want an object to hold it? Otherwise you'll get into a situation where you have absolute indexes throughout your code.
If, for some reason you don't care about the actual columns (maybe this is for a reporting tool?), then define an object that knows about the key columns and holds the rest as an ArrayList.