Browse files

Updated module documentation for B-M.

  • Loading branch information...
1 parent 8483c15 commit 9f1c6faeb1085af889b168dd736dac6cc1720da1 @bjornbm committed Apr 9, 2011
Showing with 5 additions and 3 deletions.
  1. +5 −3 Data/Random/Normal.hs
@@ -35,9 +35,11 @@ With custom mean and standard deviation:
> sample <- normalIO' (mean,sigma)
> samples <- normalsIO' (mean,sigma)
-Internally the library uses the Central Limit Theorem to approximate
-normally distributed values from multiple uniformly distributed
-random values.
+Internally the library uses the Box-Müller method to generate
+normally distributed values from uniformly distributed random values.
+If more than one sample is needed taking samples off an infinite
+list (created by e.g. 'normals') will be roughly twice as efficient
+as repetedly generating individual samples with e.g. 'normal'.

0 comments on commit 9f1c6fa

Please sign in to comment.