http://www.paranormal-entertainment.com/idr/blog/posts/2009-09-23T01:19:45Z-Floating-point_equality_for_the_lose/ blogtest IanRomanick http://www.paranormal-entertainment.com/idr/blog/posts/2009-09-23T01:19:45Z-Floating-point_equality_for_the_lose/comment_1/ http://www.paranormal-entertainment.com/idr/blog/posts/2009-09-23T01:19:45Z-Floating-point_equality_for_the_lose/comment_1/ Thu, 24 Sep 2009 11:34:17 -0700 2009-09-24T18:34:18Z It turns out that <code>0.0</code> and <code>{0.0}</code> are different as constants in fragment and vertex programs. The former expands to a vector of <code>{0.0, 0.0, 0.0, 0.0}</code> as you would expect. The later, which is what I used in my program, expands to <code>{0.0, 0.0, 0.0, 1.0}</code>. This follows the same rules as, for example, <code>glVertex3f</code>, <code>glColor3f</code>, etc., so it's not too insane. Ah... the joys of coding in assembly.