MathOverflow is a question and answer site for professional mathematicians. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Is it possible in some way, to use Tunnells's theorem to determine how long it will take a computer, to determine whether a number n, is a possible area of a rational right triangle?

share|cite|improve this question
Using BSD or unconditionally? – Alex B. Dec 7 '10 at 13:32
Assuming BSD, the answer is YES! You have to compute the number of times 2 positive ternary quadratic forms represent a number. I think roughly speaking this takes O(n^{3/2}). There was a project in a SAGE days to compute the q-expansion of the ternary form from Tunnell up to 10^{12} (the only reference I found is may be there are newer ones) – A. Pacetti Dec 7 '10 at 17:29
Thanks a lot! Think that is all I need.. – Nøhr Dec 8 '10 at 12:03
I have explained in how BSD and Tunnell's theorem imply that an integer $\alpha=jn$, with $j\in\{1,2\}$ and $n$ odd, is congruent if and only if $c_j(n)=0$, where $c_j(n)$ is the coefficient of $T^n$ in the formal series $g(T)\theta_j(T)$, with $$ g(T)=T\prod_{n=1}^{+\infty}(1-T^{8n})(1-T^{16n})\quad\hbox{and}\quad \theta_j(T)=1+2\sum_{n=1}^{+\infty}T^{2jn^2}. $$ – Chandan Singh Dalawat Jan 25 '11 at 14:19
@Pacetti: using Chandan's description we have a $\tilde{O}(n)$ algorithm - compute the product mod $T^{n+1}$ using FFT. – Dror Speiser Feb 26 '11 at 15:24

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Browse other questions tagged or ask your own question.