Errata to _Graphics Gems V_, first edition, edited by Alan Paeth
(awpaeth@okanagan.bc.ca), Academic Press 1995.  Code available online in
http://www.graphicsgems.org/

compiled by Eric Haines (erich@acm.org) from author and reader contributions

version 1.13
date:  9/11/2009

-----

Errors in the text:

The following proof changes might not appear in the book's 1st printing but
are correct on the floppy disk and FTP mirror versions:

p. 44, Girard's formula is listed as "A+B+C-2PI" should be "A+B+C-PI".
        The same mistake occurs in equation (1) on the same page,
        "2(n-2)PI" should be "(n-2)PI".

p. 76, (b) Parallel Lines: it says the quadratic equation will either have
        coefficients of A, D, and F or coefficients of B, E and F. However,
        if you multiply out [x y 1] M [x y 1]' where M is either of the two
        matrices listed under Eqn 20, you get coefficients of A, 2*D, and F
        or B, 2*E, and F. (See related error in code, below.) [thanks to
        Sarah Edwards]

p. 85, bottom (code line) now reads:

                 ...    if ((t = a - b) < 0) {a -= t; b += t; } }

                                 ^   ^          ^       ^
      ('a','b' and '+','-' were transposed)

p. 86, top (code):
        ...  + 16*d)/ ...
               ^^ replaces the ' 4' presently there

p. 153: formulas (ix) a and b are correct, but they would be better if they
    were written as:

    (a) B_k^n(t) = Sigma (-1)^j-k Binomial(j,k) Binomial(n,j) t^j

    (b) Similarly, replace Binomial(n,i,j) Binomial(n-i-j, k-i, l-j) by
        Binomial(k,i) Binomial(l,j) Binomial(n,k,l)

p. 323: no cedilla in "Francois" in author's name (cp. p 405, bottom)

p. 327: Figure 5b has an expansion of 3 vertical lines. In 5a these are 5
        pixels high, in Figure 5b they incorrectly expand to 21 pixels high;
        these should be 20 pixels high.

p. 394: Atul Narkhede's email address is now atul@yamuna.asd.sgi.com

p. 382 on: Daniel Green is now Melinda Green, melinda@superliminal.com; Don
        Hatch's email address is now hatch@hadron.org

-----

The following are errors in the book's code listings (corrected in the online
code at http://www.graphicsgems.org/).  Note that some of the
code listings online are different in minor and major ways from the code in
the book.


Serious errors (ones your compiler cannot or may not catch):

ch1-4/rat.c - page 29, line 42, change

    carry = t3&0xFFFF; lohi = (t3<<16)&0xFFFF;
        to
    carry = (t3>>16)&0xFFFF; lohi = t3&0xFFFF;

ch1-4/rat.c - page 31, line 11, change

        ck = ck<<n + num/ak1;
        to
        ck = (ck<<n) + (num/ak1);

ch2-6/conmat.c - line 397, "if (d <= 0.0)" should be "if (d < 0.0)"
        lines 502-503, P2.X gets set twice; the second time should be
        setting P2.Y.
        Line 524 should be PolyCoef[1] = 2*TempConic.D;
        Line 544 should be PolyCoef[1] = 2*TempConic.E;
        (see related text fixes earlier.)
        [thanks to Sarah Edwards for all conmat.c fixes]

ch3-4/xcoord.c - page 118, line 15, change

        sumxy += pt[k].y*zeta[k];
        to:
        sumdy += pt[k].y*zeta[k];
        [thanks to Urs Eisenmann]

ch3-4/xcoord.c - page 120, at top, in SecondOrderLeastSquares, old code:

        for(k =0; k < npoints; k++) {
            sumd   += eta[k];
            sumdx  += pt[k].x *eta[k];
            sumdx2 += px2 *eta[k];	<-- px2 computation missing!
            sumdy  += pt[k].y *eta[k];
            sumdy2 += py2 *zeta[k];	<-- py2 computation missing & should be eta!
            sumdxy += pt[k].x *pt[k].y *eta[k];
            }

	to:

        for(k =0; k < npoints; k++) {
            px2 = SQR(pt[k].x);
            py2 = SQR(pt[k].y);
            sumd   += eta[k];
            sumdx  += pt[k].x *eta[k];
            sumdx2 += px2 *eta[k];
            sumdy  += pt[k].y *eta[k];
            sumdy2 += py2 *eta[k];
            sumdxy += pt[k].x *pt[k].y *eta[k];
        }
        [thanks to Philip Oberhammer]

ch3-6/axd.* - these files have been reworked to correctly match the macro
        library that they use. See the online version.

ch4-8/qbezier.c - line 76, change "if (k = 0)" to "if (k == 0)".

ch6-4/chainCode.C - line 70, add "int trueLength = strlen(code);" and
        change all "length" to "trueLength" in postProcess(). The old code loops
        through the whole allocated string instead of just the part with data.

ch7-1/wave.c - remove the definition "double a" from each of the routines
        Rwave, Twave, and Swave.  The variable "a" is passed in and so should
        not be defined here.

        The wave.c code was found to have numerous errors by George Ogata.
        Rather than list all the fixes here, simply see the new file in the
        archive. George included an OpenGL test program for using this function,
        which can be put at the end of wave.c for testing.

#include "gl.h"
#include "glut.h"
#include <stdlib.h>

#define XMIN  0.0f
#define XMAX 20.0f
#define NUM_SAMPLES 1000

// arguments to the wave functions -- play with these
#define s 0.0f
#define f 1.0f
#define a 1.0f

void draw(void) {
  float t = glutGet(GLUT_ELAPSED_TIME) / 1000.0;
  glClear(GL_COLOR_BUFFER_BIT);
  glPushMatrix(); {
    glTranslatef(-1, 0, 0);
    glScalef(2/(XMAX-XMIN), 0.4, 1);
    //
    glTranslatef(0, 1, 0);
    glBegin(GL_LINE_STRIP);
    for (float x = XMIN; x < XMAX; x += (XMAX-XMIN)/NUM_SAMPLES)
      glVertex2f(x, Rwave(x+t*5, s, f, a));
    glEnd();
    //
    glTranslatef(0, -1.5, 0);
    glBegin(GL_LINE_STRIP);
    for (float x = XMIN; x < XMAX; x += (XMAX-XMIN)/NUM_SAMPLES)
      glVertex2f(x, Twave(x+t*5, s, f, a));
    glEnd();
    //
    glTranslatef(0, -1.5, 0);
    glBegin(GL_LINE_STRIP);
    for (float x = XMIN; x < XMAX; x += (XMAX-XMIN)/NUM_SAMPLES)
      glVertex2f(x, Swave(x+t*5, s, f, a));
    glEnd();
    //
  } glPopMatrix();
  glutSwapBuffers();
}

void quit(unsigned char key, int x, int y) {
  exit(0);
}

int main(int argc, char **argv) {
  glutInit(&argc, argv);
  glutInitWindowSize(800, 600);
  glutInitDisplayMode(GLUT_DOUBLE);
  glutCreateWindow("");
  glutIdleFunc(draw);
  glutKeyboardFunc(quit);
  glutMainLoop();
  return 0;
}

ch7-5/misc.c - if you do not have the log2() function in your compiler, use:
        #define log2(x) (log((double)x)/log(2.0))

-----

Syntax errors (ones that are not usually harmful, or are easily caught):

There are various "lint" type errors in the text's and diskette's code which
have been cleaned up in the FTP distribution.  The only serious changes were
to the axd.c code in ch3-6, as the code was out of sync with the macros it
used from ch7-7/mactbox.  The corrected code is in the FTP distribution.

-----

The following are typographical errors in the comments:

[none so far]

-----

END