Path: utzoo!attcan!uunet!wuarchive!julius.cs.uiuc.edu!apple!hercules!fernwood!oracle!news
From: dpawson@oracle.com
Newsgroups: comp.graphics
Subject: Graphics Gems Bug?
Message-ID: <1990Oct3.163001.7150@oracle.com>
Date: 3 Oct 90 16:30:01 GMT
Reply-To: dpawson@oracle.com ()
Organization: Oracle Corporation, Belmont, CA
Lines: 17

Pardon me if this has been mentioned before but. . .
Last night I was playing with the matrix inversion routine from Glassner's
Graphics Gems, and I can't help but think there is a bug in the sample code.
The routine first calls
   adjoint(in, out);
Then it finds the determinant to divide all elements of the adjoint matrix
by.  But it does this:
   det=Det4x4(out); /* Det4x4 may be wrong, but you know what I mean :-) */
Don't we instead want
   det=Det4x4(in);
?
I am rather certain we do, given that 1) it makes sense that way, and even more
importantly 2) my program works if I do it that way :-)

      	       	     	      	       	  Thanks,
      	       	     	      	       	     Dave Pawson
      	       	     	      	       	     dpawson@oracle.com