Modular arithmetic on elements of small norm in quadratic fields (Q697632)

A well-known problem when working with units in real quadratic fields is that of coefficient explosion of the fundamental unit, which makes explicit work with the unit impossible even for moderate discriminants. Recently, techniques that work on different representations (the unit is expressed as a product of smaller numbers) have been utilized to overcome this. The main method developed in this article is an efficient algorithm allowing the computation of the coefficients of a quadratic number in product representation modulo some integer. It works by successively modifying the compact representation introduced in [\textit{J. Buchmann, C. Thiel} and \textit{H. Williams} in: Bosma, Wieb (ed.) et al., Computational algebra and number theory. Sydney, 1992. Dordrecht: Kluwer Academic Publishers, Math. Appl., Dordr. 325, 159-185 (1995; Zbl 0833.11060)] to avoid denominators not coprime to the modulus. The authors then apply this to the norm equations arising from the investigation of Diophantine \(m\)-tuples (a set of positive integers \(\{a_1, \dots, a_m\}\) is a Diophantine \(m\)-tuple iff any number \(a_ia_j+1\) is a perfect square for \(i\neq j\)) to improve the classification result [\textit{A. Dujella} and \textit{A. Pethö}, Q. J. Math., Oxf. II. Ser. 49, 291-306 (1998; Zbl 0911.11018)]: If \(\{1, 3, c\}\) is a Diophantine triple then one gets \(c = c_k\) with an explicitly given sequence \(c_k\) and \(k>100\) or \(k=1\). The negative result for \(k=37\) depends on the ERH for the computation of the class and unit group in the associated quadratic field.