A news post by Raymond Toy who contradicts the claim that CMUCL uses a multiplication algorithm more efficient that O(n^2): No, CMUCL doesn't use Karatsuba.
Found this page when looking for a Spigot algorithm for pi, but has also an article about using formal methods in the development of the Transputer Floating Point Unit.
Cool page about (semi)numerical algorithm, including a book in progress, an open source library and a number of interesting links. Also has a mirror of hakmem. The author has also published a book about pi.
Contains references to B. Smiths ZERPOL and Adam's stopping criterion. SMITH, B.T. ZERPOL, a zero finding algorithm for polynomials using Laguerre's method. Proc. 1967 Army Numerical Analysis Conference, Madison, Wis., May 1967 (Rep. 67-3, US Army Res. O