NAME

Math::VectorReal - Module to handle 3D Vector Mathematics

SYNOPSIS

#!/usr/bin/perl
use Math::VectorReal;

$a = vector( 1, 2, .5 );
print "Vector as string (MatrixReal default format)\n\$a => ", $a;

print  $a->stringify("Formated Output   \$a => { %g, %g, %g }\n");

# I hate newline in the default output format (defined as MatrixReal)
$Math::VectorReal::FORMAT = "[ %.5f %.5f %.5f ]";
print "Modified default output format   \$a => $a\n";

print 'length     => ', $a->length, "\n";
print 'normalised => ', $a->norm, "\n";

use Math::VectorReal qw(:all);  # Include O X Y Z axis constant vectors
print 'string concat   $a."**" = ', $a."**", "\n";
print 'vector constant    X    = ',   X,    "\n";
print 'subtraction     $a - Z  = ', $a - Z, "\n";
print 'scalar divide   $a / 3  = ', $a / 3, "\n";
print 'dot product     $a . Y  = ', $a . Y, "\n";
print 'cross product   $a x Y  = ', $a x Y, "\n";

print "Plane containing points X, \$a, Z (in anti-clockwise order)\n";
($n,$d) = plane( X, $a, Z ); # return normal and disance from O
print '      normal      =    $n     = ', $n, "\n";
print '  disance from O  =    $d     = ', $d, "\n";
print ' Y axis intersect = $d/($n.Y) = ', $d/($n.Y), "\n";

print "VectorReal and MatrixReal interaction\n\n";
use Math::MatrixReal;  # Not required for pure vector math as above

$r = $a->vector2matrix_row;  # convert to MatrixReal Row Vector
$c = $a->vector2matrix_col;  # convert to MatrixReal Column Vector
print 'Vector as a MatrixReal Row $r (vector -> matrix) => ', "\n", $r;
print 'Vector as a MatrixReal Col $c (vector -> matrix) => ', "\n", $c;

$nx = $a->norm;   $ny = $nx x Z;  $nz = $nx x $ny; # orthogonal vectors
$R = vector_matrix( $nx, $ny, $nz );   # make the rotation matrix
print 'Rotation Matrix from 3 Vectors   $R   => ',"\n", $R, "\n";

print "Extract the Y row from the matrix as a VectorReal..\n";
print '$R->matrix_row2vector(1) => ', $R->matrix_row2vector(1), "\n";

print "Rotate a vector with above rotation matrix\n";
print '$a * $R (vector -> vector)',"\n", $a * $R, "\n";

print "Rotate a MatrixReal column (post multiply)...\n";
print "(NB: matrix must be transposed (~) to match column format)\n";
print '~$R * $c (col_matrix -> col_matrix) =>',"\n", ~$R * $c, "\n";

DESCRIPTION

The Math::VectorReal package defines a 3D mathematical "vector", in a way that is compatible with the previous CPAN module Math::MatrixReal. However it provides a more vector oriented set of mathematical functions and overload operators, to the MatrixReal package. For example the normal perl string functions "x" and "." have been overloaded to allow vector cross and dot product operations. Vector math formula thus looks like vector math formula in perl programs using this package.

While this package is compatible with Math::MatrixReal, you DO NOT need to have that package to perform purely vector orientated calculations. You will need it however if you wish to do matrix operations with these vectors. The interface has been designed with this package flexibility in mind.

The vectors are defined in the same way as a "row" Math::MatrixReal matrix, instead of that packages choice of "column" definition for vector operations. Such vectors are multiplied to matices with the vector on the left and the matrix on the right. EG: v * M -> 'v

Not only is this the way I prefer to handle vectors, but it is the way most graphics books use vectors. As a bonus it results in no overload conflicts between this package and that of Math::MatrixReal, (the left objects overload operator is called to do the mathematics). It also is a lot simpler than MatrixReal column vector methods, which were designed for equation solving rather than 3D geometry operations.

The vector_matrix() function provided, simplifies the creation a MatrixReal object from 3 (usually orthogonal) vectors. This with its vector orientated math operators makes it very easy to define orthogonal rotation matrices from Math::VectorReal objects. See a rough example in the synopsis above, or in the file "matrix_test" in the packages source.

NOTE: the 6th element the Math::MatrixReal array object is used to hold the length of the vector so that it can be re-used without needing to be re-calculated all the time. This means the expensive sqrt() function, need not be called unless nessary. This usage should not effect the direct use of these objects in the Math::MatrixReal functions.

CONSTANTS

Four constant vectors are available for export (using an ":all" tag). these are

0 = [ 0 0 0 ]   the zero vector or origin
X = [ 1 0 0 ]   |
Y = [ 0 1 0 ]    > Unit axis vectors
Z = [ 0 0 1 ]   |

CONSTRUCTORS

new(x,y,z)

Create a new vector with the values of x, y, z returning the appropriate object.

vector(x,y,z)

As new but is a exported function which does not require a package reference to create a Math::VectorReal object.

clone()

Return a completely new copy of the referring Math::VectorReal object.

METHODS

array()

Return the x,y,z elements of the referring vector are an array of values.

x()

Return the x element of the referring vector.

y()

Return the y element of the referring vector.

z()

Return the z element of the referring vector.

stringify( [ FORMAT ] )

Return the referring verctor as a string. The FORMAT if given is used to sprintf format the vector. This is used for all VectorReal to String conversions.

By default this format is the same as it would be for a Math::MatrixReal object, "[ %#19.12E %#19.12E %#19.12E ]\n". Note that this includes a newline character!.

However unlike Math::MatrixReal you can assign a new default sprintf format by assigning it to the packages $FORMAT variable. For Example

$Math::VectorReal::FORMAT = "{ %g, %g, %g }"

Which is a good format to output vectors for use by the POVray (Persistance of Vision Raytracer) program.

length()

Return the length of the given vector. As a side effect the length is saved into that vectors object to avoid the use of the expensive sqrt() function.

norm()

Normalise the Vector. That is scalar divide the vector by its length, so that it becomes of length one. Normal vectors are commonly use to define directions, without scale, or orientation of a 3 dimensional plane.

plane( v1, v2, v3 )

Given three points defined counter clockwise on a plane, return an array in which the first element is the planes normal unit vector, and the second its distance from the origin, along that vector. NOTE: the distance may be negitive, in which case the origon is above the defined plane in 3d space.

vector_matrix( nx, ny, nz )

Given the new location for the X, Y and Z vectors, concatanate them together (row wise) to create a Math::MatrixReal translation matrix. For example if the 3 vectors are othogonal to each other, the matrix created will be a rotation matrix to rotate the X, Y and Z axis to the given vectors. See above for an example.

VECTOR/MATRIX CONVERSION

The following functions provide links between the Math::VectorReal and Math::MatrixReal packages.

NOTE: While this package is closely related to Math::MatrixReal, it does NOT require that that package to be installed unless you actually want to perform matrix operations.

Also the overload operations will automatically handle vector/matrix mathematics (See below).

Vector to Matrix Conversion

vector2matrix_row( [CLASS] )

vector2matrix_col( [CLASS] )

Convert Math::VectorReal objects to a Math::MatrixReal objects. Optional argument defines the object class to be returned (defaults to Math::MatrixReal).

Note that as a Math::VectorReal is internally equivelent to a Math::MatrixReal row matrix, vector2matrix_row is essentually just a bless operation, which is NOT required to use with Math::MatrixReal functions.

The vector2matrix_col performs the required transpose to convert the Math::VectorReal object into a Math::MatrixReal version of a vector (a column matrix).

Matrix to Vector Conversion

matrix_row2vector( [ROW] )

matrix_col2vector( [COLUMN] )

When referred to by a Math::MatrixReal object, extracts the vector from the matrix. the optional argument defines which row or column of the matrix is to be extracted as a Math::VectorReal vector.

OPERATOR OVERLOADING

Overload operations are provided to perform the usual string conversion, addition, subtraction, unary minus, scalar multiplation & division. On top of this however the multiply have been expanded to look for and execute MatrixReal multiplation.

The Main purpose of this package however was to provide the special vector product operations: dot product "." and cross product "x". In perl these operations are normally used for string operations, but if either argument is a VectorReal object, the operation will attempt the approprate vector math operation instead.

Note however that if one side of the dot "." operator is already a string, then the vector will be converted to a sting and a string concatantion will be performed. The cross operator "x" will just croak() as it is non-sensical to either repeat the string conversion of a vector, OR to repeat a string, vector, times!

Overloaded operator summery... neg unary minus - multiply vector by -1 "" automatic string conversion using stringify() function + vector addition - vector subtraction / scalar division (left argument must be the vector) * scalar multiplication OR MatrixReal multiplication x vector/cross product of two vectors . dot product of two vectors OR vector/string concatanation

Posible future addition '~' to transpose a VectorReal into a MatrixReal column vector (as per that operator on MatrixReal objects). It was not added as it just did not seem to be needed.

SEE ALSO

The Math::MatrixReal CPAN Module by Steffen Beyer and the Math::MatrixReal-Ext1 CPAN extension by Mike South

AUTHOR

Anthony Thyssen <anthony@cit.gu.edu.au>

COPYRIGHT

Copyright (c) 2001 Anthony Thyssen. All rights reserved. This program is free software; you can redistribute it and/or modify it under the same terms as Perl itself. I would appreciate any suggestions however.

cpanm PerlMol

perl -MCPAN -e shell
install PerlMol