The product of like-indexed terms in binary recurrences (Q1864854)

In recent work of \textit{L. Szalay} [On the Diophantine equation \((2^n-1)(3^n-1)= x^2\), Publ. Math. 57, 1-9 (2000; Zbl 0961.11013)] and \textit{L. Hajdu} and \textit{L. Szalay} [On the Diophantine equations \((2^n-1)(6^n-1)= x^2\) and \((a^n-1)(a^{kn}-1)= x^2\), Period. Math. Hung. 40, 141-145 (2000; Zbl 0973.11015)] Diophantine equations of type \[ (a^k-1)(b^k-1)= x^n, \tag{1} \] for given distinct integers \(a>1\) and \(b>1\), are completely solved for a few pairs \((a,b)\) and \(n=2\). The authors of the paper under review consider (1) in integers \((k,x,n)\) with \(n>1\). They show that these equations have only finitely many solutions. This result is in fact a corollary of a more general finiteness theorem on equations of type \[ (c_1a_1^k+ d_1b_1^k) (c_2a_2^k+ d_2b_2^k)= ex^n, \] a result that is ineffective. However, for \(n=2\) and almost all pairs \((a,b)\) with \(2\leq b< a\leq 100\) (actually for more than 98\% of them) the authors completely solve equation (1). Finally it is shown that the \(abc\)-conjecture implies much stronger results on this type of equation.