References (Score 5, Informative)

I was about to post something longer, but my computer deleted it at just the right moment. Summary: (a) The origin of this article was an interview given at an algebra colloquium in Buenos Aires a few weeks ago. Its appearance in Scientific American was a little unexpected (to me). I should be able to post a preprint soon. (b) In my colloquium talk, I made sure to mention that I had just come across a precedent in the literature: W. F. Galway, Dissecting a sieve to cut its need for space. In Algorithmic number theory (Leiden, 2000), vol 1838 of Lecture Notes in Comput. Sci., pages 297-312. Springer, Berlin, 2000. In both cases, we are talking about space essentially O(N^(1/3)) and time essentially O(N). Galway improves on the Atkin-Bernstein sieve, which is specifically about producing lists of primes. The sieve of Eratosthenes can be used for more general purposes, e.g., producing lists of factorizations or computing tables of values of arithmetic functions that depend on factorization. I actually got interested in the problem while using a sieve to produce successive values of mu(n), where mu is the Moebius function. As far as I can tell, there's a basic underlying idea being used in both cases, viz. Diophantine approximation. In brief, if you are finding primes (or what have you) in an interval [N,N+N^(1/3)], you do not have to sieve by every m=sqrt(N); you can tell in advance which integers m will have no multiples in that interval. This is all more or less orthogonal to [http://sweet.ua.pt/tos/software/prime_sieve.html]. Indeed what I do and what Oliveira e Silva does can very likely be combined. Best Harald Helfgott