On \(k\)-caps in \(\mathrm{PG}(n, q)\), with \(q\) even and \(n \geq 3\) (Q1709549)

A \(k\)-arc of \(\mathrm{PG}(2, q)\) is a set of \(k\) points, no three of which are collinear; a \(k\)-cap of \(\mathrm{PG}(n, q)\), \(n \geq 3\), is a set of \(k\) points, no three of which are collinear. A \(k\)-arc or \(k\)-cap is complete if it is not contained in a \((k+1)\)-arc or \((k+1)\)-cap. The largest value of \(k\) for which a \(k\)-arc of \(\mathrm{PG}(2, q)\), or a \(k\)-cap of \(\mathrm{PG}(n, q)\) with \(n \geq 3\), exists is denoted by \(m_2(n, q)\). The size of the second largest complete \(k\)-arc of \(\mathrm{PG}(2, q)\) or \(k\)-cap of \(\mathrm{PG}(n, q)\), \(n \geq 3\), is denoted by \(m^{\prime}_2(n, q)\). In [Geom. Dedicata 74, No. 1, 91--94 (1999; Zbl 0920.51004)], \textit{J. Chao} proves that for \(q\geq 8\) even \(m^{\prime}_2(3, q) \leq q^2 - q + 5\). In the paper under review, the author improves the above bound to \(q^2 - (\sqrt{5}-1)q + 5\) and new bounds for \(m_2(n, q)\), \(q\) even, \(q \geq 8\) and \(n \geq 4\), are obtained. Using [\textit{L. Storme} and \textit{T. Szönyi}, Geom. Dedicata 45, No. 2, 163--169 (1993; Zbl 0770.51015)], the bound for \(m^{\prime}_2(3, q)\), \(q\geq 2048\), can be improved to \(q^2 - 2q + 3\sqrt{q} + 2\). As the author remarks, the obtained bound is worse than the one given in [\textit{J.-M. Cao} and \textit{L. Ou}, Discrete Math. 326, 61--65 (2014; Zbl 1290.51004)], namely \(q^2 - 2q + 8\) (\(q\) even and \(q \geq 128\)). Unfortunately, it seems that some of the arguments in [loc. cit., Section 1.3] are not correct.