Try this: %!PS-Adobe-3.0 %%BoundingBox: 0 0 612 792 %%Orientagion: Portrait %%Pages: 1 %%EndComments %%Page: 1 1 %%BeginPageSetup %%EndPageSetup <</PageSize [612 792]>> setpagedevice 0 setlinewidth 0 setlinecap 8.5 36 mul 11 36 mul translate 100 100 scale /c1 5 sqrt 1 sub 4 div def /c2 5 sqrt 1 add 4 div def /s1 10 20 sqrt add sqrt 4 div def /s2 10 20 sqrt sub sqrt 4 div def /-c1 c1 neg def /-c2 c2 neg def /-s1 s1 neg def /-s2 s2 neg def .8 setgray /p2 {-1 0 moveto -c1 s1 lineto c2 s2 lineto c2 -s2 lineto -c1 -s1 lineto fill} bind readonly def /2c2 c2 2 mul def gsave 2c2 neg 0 translate p2 0 setgray -c1 s1 moveto -c1 -s1 lineto stroke grestore gsave 2c2 -c1 mul 2c2 s1 mul translate p2 grestore gsave 2c2 c2 mul 2c2 s2 mul translate p2 grestore gsave 2c2 c2 mul 2c2 -s2 mul translate p2 grestore gsave 2c2 -c1 mul 2c2 -s1 mul translate p2 grestore 1 0 moveto c1 s1 lineto -c2 s2 lineto -c2 -s2 lineto c1 -s1 lineto fill 0 setgray c1 s1 moveto c1 -s1 lineto stroke showpage On Friday 20 August 2010 08:45:36 Bill Gosper wrote:
Apologies. I had tried superposing the segments with Windows Paint, which has a mode where white <-> transparent. But for some bizarre reason, it enforced opaque white on this image, and I couldn't test it. (One *very* frustrating hour later:) I tried rewriting it at various pixel depths and as a gif. Paint not only refused to treat white as transparent, it even disabled the feature in the toolbar, and left it disabled when I loaded a file on which it formerly worked! So I tried "physically" moving a rectangle containing the longer line onto the shorter, and the black part became transparent!! Paint has gone criminally insane. If this were something important, I would be, too. Anyway, I just manually fudged the picture closer to equality, and I still see the illusion. Feel free to repair it with some sane software, and see if it still works for you. --rwg
On Thu, Aug 19, 2010 at 8:44 PM, Bill Thurston <wpt4@cornell.edu> wrote:
What is your experimental method, guys? I copied the image into Adobe Illustrator and used the measuring tool to check.
The leftward segment measured 4.05", and the rightward segment measured 4.17". A 3% difference. Theoretical equality != actual equality.
My experimental method was somewhat cruder: I held up a piece of paper next to one line on my screen, made two marks, then slid the paper over. Going back and forth a few times convinced me that it wasn't just pencil mark error.
--Joshua
_______________________________________________ math-fun mailing list math-fun@mailman.xmission.com http://mailman.xmission.com/cgi-bin/mailman/listinfo/math-fun