3

Whenever I play a playlist of music using VLC (possibly other software too), I notice that some songs never get played while others get played repeatedly (even for a playlist of just 8 songs).

I know from my days as a computer science undergrad student that practical randomness is often good enough so we don't bother with true randomness for many applications. But I'd like to understand this better.

  • why is real randomness not worth the effort?
  • does pseudo-randomness not follow a uniform distribution?
  • are software developers perfectly aware of behavior like I see in VLC and just see a tiny market segment as caring about it?

EDIT: actually, my VLC issue might be because of a broken link! But I'm sure I've seen this behavior with other music players like Apple Music. So apologies to the good folks who develop VLC.

Sridhar Sarnobat
  • 173
  • 6

1 Answers1

5

This is just speculation, but perhaps what VLC is attempting is to simulate... perfect randomness. That is, each song is picked uniformly at random, independently of previous songs. According to the coupon collector problem, if you have $n$ songs in your list and want to hear all of them, you will have to wait for roughly $n\ln n$ songs to be played. By that time, the average song will have been heard $\ln n$ times.

The problem seems to be not in VLC, but rather in the random experiment that VLC is trying to simulate. VLC is not trying to shuffle your playlist and play it through. Rather, it seems to just pick a new random song each time, regardless of the past.

Yuval Filmus
  • 280,205
  • 27
  • 317
  • 514