From 34b855b078f316b3c23a2483c0d45401cfe10a96 Mon Sep 17 00:00:00 2001 From: stevenj Date: Wed, 5 Sep 2007 17:25:43 -0400 Subject: [PATCH] tweak darcs-hash:20070905212543-c8de0-a6f8bacf2db3c0b3ae79cd0f06ad946f2afdb898.gz --- util/sobolseq.c | 15 ++++++++------- 1 file changed, 8 insertions(+), 7 deletions(-) diff --git a/util/sobolseq.c b/util/sobolseq.c index 12c3d5d..7698d03 100644 --- a/util/sobolseq.c +++ b/util/sobolseq.c @@ -27,13 +27,14 @@ as modified by: S. Joe and F. Y. Kuo, ACM Trans. Math. Soft 29 (1), 49-57 (2003). - Note that the code below was written without even looking at the Fortran - code, which is only semi-free (being under the restrictive ACM copyright - terms). Then I went to the Fortran code and took out the table of - primitive polynomials and starting direction #'s ... since this is just - a table of numbers generated by a deterministic algorithm, it is not - copyrightable. (Obviously, the format of these tables necessitated - some slight modifications to the code.) + Note that the code below was written without even looking at the + Fortran code from the TOMS paper, which is only semi-free (being + under the restrictive ACM copyright terms). Then I went to the + Fortran code and took out the table of primitive polynomials and + starting direction #'s ... since this is just a table of numbers + generated by a deterministic algorithm, it is not copyrightable. + (Obviously, the format of these tables then necessitated some + slight modifications to the code.) For the test integral of Joe and Kuo (see the main() program below), I get exactly the same results for integrals up to 1111 -- 2.30.2