What Can Be Computed in Algebraic Geometry

This paper evolved from a long series of discussions between the two authors, going backto around 1980, on the problems of mak ing effective computations in algebraic geometry, and it tookmore definite shape in a survey talkgiven by the second author at a conference on Computer Algebra in 1984. The goal at that time was to bring together the perspectives of theoretical computer scientists and of working algebraic geometers, while laying out what we considered to be the main computational problems and bounds on their complexity. Only part of the talkwas written down and since that time there has been a good deal of progress. However, the material that was written up may still serve as a useful introduction to some of the ideas and estimates used in this field (at least the editors of this volume thinkso), even though most of the results included here are either published elsewhere, or exist as “folk-theorems” by now. The article has four sections. The first two parts are concerned with