// G-Force DeltaField
// Name: HAL_Mystery_Box
// Created by: Howard A. Landman, 26 July 2001
// e-mail: howard@polyamory.org -or- howard.landman@vitesse.com
// homepage: http://www.polyamory.org/~howard/

// Dedicated to Mickey Hart

// This one jumps through flaming hoops to invert the radius while still
// remaining completely true to the rectangular boundary.  Then it throws
// a touch of chaos and rotation into the inversion just to pretty things up.

// The "0.000001" terms are to prevent divide-by-zero problems.
// Try taking them out and you'll see what I mean.

Aspc=0,

A0="rnd(0.2)-rnd(0.2)",    // speed of rotation

A1="1+rnd(0.2)-rnd(0.1)",  // degree of chaos, 0.9 to 1.2
                           // When this is < 1 we get a mystery box in the middle.
                           // When it's > 1 we get more of a rectangular corridor with no blatant box.

// Calculate maximum radius at this angle without going outside of the screen
D0="sgn(abs(sin(theta))-abs(cos(theta)))",  // positive if sine is bigger than cosine
D1="(d0+1)*0.5",  // 1 if sine is bigger than cosine, 0 otherwise
D2="d1/(abs(sin(theta))+0.000001) + (1-d1)/(abs(cos(theta))+0.000001)", // max radius at our T

// Calculate maximum radius at source angle without going outside of the screen
D3="sgn(abs(sin(theta+a0))-abs(cos(theta+a0)))",  // positive if sine is bigger than cosine
D4="(d3+1)*0.5",  // 1 if sine is bigger than cosine, 0 otherwise
D5="d4/(abs(sin(theta+a0))+0.000001) + (1-d4)/(abs(cos(theta+a0))+0.000001)", // max radius at srcT

//srcR="d5*(1 - r/d2)",         // This would be the simple inversion
                                // "maxR - R", *not* "1/R" !)
srcR="d5*a1*((1 - r/d2)-0.5)",  // inversion with a bit of chaos
srcT="theta + a0",              // some rotation too

// Actually working out the right math here was fun -
// I haven't had to use secants in a while.

Vers=200