When I read this, I was immediately suspicious, partly because the claimed accuracy sounded incredible, and partly because the construction involves an arbitrary choice (which, so to speak, "contradicts" the particularity of the error-estimate given, which is probably just for one value that that choice).
So I started to try to disprove it, but since I've now verified quite a bit of it, I'm beginning to change my mind...
On 20 Jun 2002, Mark wrote:
Come on Mark, tell us your full name, and how you came on this remarkable idea!
> Just wanted to share this, I worked it out a couple of years ago and > was surprised by the accuracy: > > Near Exact Trisection: > 1. Start with an unknown angle <90 deg., label the vertex A. > 2. Draw an arc with origin at A crossing both lines of the angle at > points B and C. > 3. Draw line BC making an isosceles triangle. > 4. Using point C as the origin, draw an arc crossing line BC and the > earlier arc somewhere between ÃÂ¼ and ÃÂ½ way between points C and B. > Label where this new arc crosses line BC point D. > Label where this new arc crosses the first arc point E. > 5. Draw line DE and extend it well past A . If line DE passes > exactly through point A (it wonÃÂt) stop, your first guess was an exact > trisection.
My first "hope" was that this would be wrong (I say "hope", because if it were wrong, I'd be absolved from checking the rest). But it's entirely correct, and already intriguing, in that it leads to a nice "neusis" construction whose simplicity rivals Archimedes'.
> 6. Extend line AC well past point A, step off 3 times length AC from > point A and label the new point F. > 7. Swing an arc of length AF with A as the origin that crosses the > extended line DE near point F. > Label the intersection G. > 8. Draw line GA and extend it to intersect the original arc from step > 2. > Label the intersection EÃÂ.
I haven't yet checked all this, but plan to. I'm already quite impressed...
> Line AEÃÂ is a good (within less than 1/1000 degree) trisection. > However this is only the start. Repeating the process from step 4 > using CEÃÂ as the arc radius results in a trisection to within 10E-11 > degrees. Each subsequent iteration improves the trisection by several > orders of magnitude.
...and plan to study this construction very carefully.