On the Diophantine equation x²+7=y^m (Q4484443)

The authors consider a conjecture posed by \textit{J. H. E. Cohn} [Acta Arith. 65, 367-381 (1993; Zbl 0795.11016)] to the effect that the exponential Diophantine equation \(x^2+7= y^m\) with \(x,y,m\in \mathbb{N}\) and \(m\geq 3\) has only the seven solutions \((x,y,m)= (1,2,3)\), \((181,32,3)\), \((3,2,4)\), \((5,2,5)\), \((181,8,5)\), \((11,2,7)\), and \((181,2,15)\). Quite a number of partial results are known about this equation, for instance, \textit{J.-L. Lesage} showed in [J. Number Theory 73, 390-425 (1998; Zbl 0923.11055)] that any solution \((x,y,m)\) must satisfy \(m\leq 6.6\cdot 10^{15}\). Restricting their attention to the case of prime exponents, the authors show that there can be no solution for prime \(m\) with \(11\leq m\leq 10^8\). It also follows from their result that for any solution not listed above, \(m\) must be prime and \(10^8< m\leq 6.6\cdot 10^{15}\). In fact, mimicking an approach by \textit{A. Kraus} [Exp. Math. 7, 1-13 (1998; Zbl 0923.11054)], the authors give a clever, rather technical criterion, by which all prime exponents \(m\) in the range \(11\leq m\leq 10^8\) can be excluded. In an epilogue the authors discuss the possibility of completely resolving Cohn's conjecture by means of their computational method. They come to the conclusion that such is not feasible unless the known lower bounds for linear forms in three (complex) logarithms can be substantially improved.