No label defined (Q5617656)

This is an excellent article, which should prove to be of great interest to number theorists who have used computers but have found the use of languages such as ALGOL and FORTRAN frustrating. The author presents a number of algorithms for prime factorization of integers, finding GCDs, continued fractions, approximate operations, etc., which are economical, though not always the fastest methods of solving such problems by machine. It is of interest to compare the method for solving a particular problem given by the author with those presented in [\textit{D. E. Knuth}, The art of computer programming. Vol. 2: Seminumerical algorithms. Reading, Mass.: Addison-Wesley (1969; Zbl 0191.18001), chap. 4] to find the most efficient algorithm for that process. It is hoped that the present article will have an influence on software designers and will persuade them to include multiprecision operations in their compilers. They must never forget that ``some operations are impossible to describe in a language like ALGOL that fails to take into account the word structure of the machine''. The introduction of PL/I is a great improvement in this respect, but much improvement is still needed. Finally, this paper contains many useful theorems which are relatively unknown, as well as material of theoretical interest, especially for primality testing. For the entire collection see Zbl 0212.30701