After recently taking it upon ourselves, Tanner, Weregoose, and myself recently made a couple of Bezier curve programs that would make a curved line of some sort based on given point inputs.

As I understand, there are two types of Bezier curves, with no names to differentiate between the two: one that becomes more accurate with smaller increments in percentages (first picture below) and the other that should become more accurate with more recursions (second picture).

...and...


Tanner and I had both tried the latter Bezier curve program, and got some weird results. After about 100 iterations, the curved line becomes more so a straight line between the first and last points. We expected it to become more accurate with more iterations, and it does, but after a certain point, it gets further from the expected curve.

Any opinions on whether the problem is with the programs, or it's because of this method of making a curve?

For the interested, this is the version I've been toying with (based around MufinMcFlufin's source here):

[.8xp]

Code:

Syntax: input:prgmBEZIER

• {0,0,0,1,1,1,1,0} for a semicircle
• {3,3,8,8,0,8,5,3} for a loop
• rand(6) for a random quadratic curve
• rand(8) for a random cubic curve
• rand(2n+2) for a random n-order curve

Ooooh, that looks very cool; I feel like trying to write one of these programs myself! And of course my initial reflex to try to perform optimization is pretty futile when the code in question is written or at least modified by Weregoose.