@robjohn: Apparently Ramanujan "derived" things, he just got rid of all of whatever scratchwork he put down and listed all the main results in bulk. (Though my reading has it he was still formally sloppy [?] and relied heavily on intuition + mental math.)
I just found an error in my calculation of the 6th order term in an asymptotic expansion; I multiplied instead of dividing. The numerical check seems to indicate that the asymptotic expansion is now correct.
@Gortaur once in a while.
be back in a bit. gotta get people up.
@J.M.: found another error via numerical checking, and now the asymptotic looks correct.
well, these things happens, but luckily you have Mathematica to double-check :) and then you needed to triple check the double checking, if I understand correctly...
@robjohn: btw. I absolutely second J. M.'s comment on meta to simply post that animated GIF. As long as they don't become the overwhelming majority, it is nice to have such an addendum.
I wouldn't mind see any packing of a square by 45-60-75 triangles. It did not seem clear to me that anyone had actually done it in an explicit sense, just in a long existence proof sense. The only answer last I looked just had some triangles packed, and it seem already challenging just to get the sides to match.
More important, I should apply this theory to treat linear equation with unbounded solution in 'Banach spaces', i.e. allowing functions f : X -> (-\infty,\infty] (sorry for weird LaTeX)
sure, I don't care much about others. My colleague tried to complain that in my example we have Ix=b+Ix so b=0 unless I told him not to subtract infinity from infinity
I can see the difficulty that people encounter using \infty as a number, but I think it is used properly in your case. It will still cause some difficulty when first looking at the problem.
Lately my answers are just not great. I guess I strive on that end-of-year period where the fun questions come from students that wonder about the axiom of choice. :D
