Integer points on elliptic curves (Q5920596)

The paper is aimed at bounding the number of integral or, more generally, of \(S\)-integral points of an elliptic curve \(E\) over a number field \(K\), where \(S\) is a finite set of the archimedian absolute values and some non-archimedian absolute values (valuations) of \(K\) (so that we obtain the integral points when \(S\) contains only the archimedian absolute values of \(K)\). According to a conjecture of Lang, for a quasi-minimal model of \(E\), the number of integral points is bounded by a constant depending only on the number field \(K\) and the rank \(r\) of \(E\) over \(K\). A weaker version of this conjecture was proved by Hindry and Silverman and made explicit by Gross and Silverman. They did not assume the curve to be quasi-minimal. The authors of the present paper improved the bound of Gross and Silverman. For in their bound there appeared the quadratic term \(r.\delta\) in the exponent of a constant in addition to the cardinality \(s\) of the set \(S\), where \(\delta\) is the number of prime ideals occurring in the denominator of the \(j\)-invariant of \(E\). In contrast the authors obtain linear expressions in \(r\) and \(\delta\) in the exponent, and their constant is smaller. And while Gross and Silverman used Roth's theorem, the authors take a completely different approach. In fact they refer to an explicit formula for the 2-torsion points of \(E\) to get the problem down to that of solving unit equations. For proving their result they need a series of lemmata involving a lot of details which are somewhat difficult to follow. But the proof is on the other hand rather elementary and somewhat computational. It would be interesting to compute and compare the two bounds in the example given in the paper [\textit{A. Pethö, H. G. Zimmer, J. Gebel} and \textit{E. Herrmann}, Computing all \(S\)-integral points on elliptic curves. Math. Proc. Camb. Phil. Soc. 127, 383--402 (1999; Zbl 0949.11033)]. However this paper, whose aim ist the determination of \(S\)-integral points of elliptic curves over number fields, is not mentioned here. We should point out that the title of the paper under consideration is somewhat misleading. Because its aim is that of bounding and not of determining the \(S\)-integral points of elliptic curves.