Arcs in \(\mathrm{AG}(2,q)\) determining few directions (Q1952258)

In the affine plane \(\mathrm{AG}(2,q)\), \(q=p^h\) an odd prime power, the defect of a \(k\)-arc \(K\) is the difference between the number of slopes of secants of \(K\) and \(k-1\). The non-existence of hyperfocused arcs, i.e. with defect \(0\), in \(\mathrm{AG}(2,q)\) was proved in [\textit{A. Bichara} and \textit{G. Korchmaros}, Ann. Discrete Math. 14, 117--121 (1982; Zbl 0493.51010)] and the classification of sharply focused arcs, i.e. with defect \(1\), was given in [\textit{A. Beutelspacher} and \textit{F. Wettl}, Des. Codes Cryptography 3, No. 2, 127--134 (1993; Zbl 0770.94004)]. Furthermore, it was proved in [\textit{T. Szönyi}, in: Combinatorics 88. Vol. 2. Rende: Mediterranean Press. 449--458 (1991; Zbl 0945.51521)] that \(k\)-arcs with defect \(2\) and \(k\geq 28\) can be uniquely extended to sharply focused arc. Moreover, if the order of the magnitude of the defect does not exceed \(\sqrt{q}/8\), then either \(k\leq (q+1)/2\) or the defect is \(q+1\). The present paper improves these results about \(k\)-arcs with defect \(2\) and gives some more embeddings results for \(k\)-arcs with defect \(3\). In particular, let \(K\) be a \(k\)-arc with defect \(\Delta\) and let \(\Gamma_{2\Delta}\) be an algebraic envelope containing the \(k\Delta\) special tangents to \(K\). When \(\Delta = 2\), the following results hold: (i) If \(p>3\) and \(k > 17\), then \(K\) is contained in a sharply focused arc of size \(k+1\). (ii) If \(p>3\) and \(12\leq k\leq 17\), then either (i) holds or the algebraic envelope \(\Gamma_{2\Delta}\) is singular. Both statements remain valid for \(p=3\) provided that \(\Gamma_{2\Delta}\) is neither elliptic nor a rational curve. When \(\Delta = 3\), it was proved in [\textit{A. Beato}, \textit{G. Faina}, and \textit{M. Giulietti}, Contrib. Discrete Math. 3, No. 1, 96--108 (2008; Zbl 1204.51010)] that if either \(k > 145\), or \(k > 108\) and \(p > 5\), then \(K\) is contained in a sharply focused \((k+2)\)-arc. This bound is improved under the hypothesis that no conic in \(\mathrm{AG}(2,q)\) meets \(K\) in at least \(8\) points. Namely: If \(p > 5\) and \(k\geq 100\) and if no conic in \(\mathrm{AG}(2,q)\) meets \(K\) in at least \(8\) points, then \(K\) is contained in a sharply focused \((k+2)\)-arc. Finally, some results concerning the existence problem of complete \((q-1)\)-arcs in \(\mathrm{PG}(2,q)\), with \(q>13\), are obtained. Namely: If \(K\) is a complete \((q-1)\)-arc in \(\mathrm{PG}(2,q)\), if \(\Gamma_6\) is non-singular and no conic meets \(K\) in at least \(8\) points, then \(q < 79\). If \(K\) is a complete \((q-1)\)-arc in \(\mathrm{PG}(2,q)\), \(\Gamma_6\) is non-singular, \(K\) is left invariant by a linear collineation of order \(2\) and no conic meets \(K\) in at least \(8\) points, then \(q \leq 47\).