Binary curves of small fixed genus and gonality with many rational points (Q2075925)

\textit{G. van der Geer} [Prog. Math. 202, 225--238 (2001; Zbl 1025.11022)] has asked ``What is the maximum number of rational points on a curve of genus \(g\) and gonality \(\gamma\) defined over \(\mathbb{F}_q\)?'' Call this quantity \(N_q(g,\gamma)\). There is a large literature devoted to studying the maximum number of points on a curve of genus \(g\) defined over \(\mathbb{F}_q\), but this paper seems to be the first significant effort to refine this question by the gonality of the curve. In this article, the authors focus on the binary field \(\mathbb{F}_2\). They determine \(N_2(g,\gamma)\) for all \(g \le 5\) and all \(\gamma\) by making use of a detailed understanding of the geometry of curves of genus at most \(5\). (In the appendix to this paper, the authors determine \(N_2(g,\gamma)\) for several additional values with \(6\le g \le 10\) using a less systematic approach.) The focus of this paper is on curves over \(\mathbb{F}_2\), but many of the results apply to curves over more general fields. The discussions of curves of genus \(4\) and \(5\) contain several interesting results about spaces of quadric hypersurfaces containing a given curve. They adapt some results that are known over algebraically closed fields to apply over more general fields. Determining the value of \(N_2(5,5)\) involves a nice mix of theoretical and computational work. The authors describe their algorithms in a way that is easy to follow for a reader who is not an expert in computer algebra. They explain a computation that proves that there does not exist a curve defined over \(\mathbb{F}_2\) of genus \(5\) and gonality \(6\). It is not difficult to show that \(N_q(g,\gamma) \le \gamma (q+1)\). The authors conjecture that for fixed \(\gamma\) and all \(g\) sufficiently large, \(N_2(g,\gamma) = 3 \gamma\). They give a simple construction that proves the conjecture in the case \(\gamma = 2\). This article has already been a jumping off point for significant further work in this area. For example, in a follow-up paper the authors use similar ideas to prove analogous classification results for curves over \(\mathbb{F}_3\) and \(\mathbb{F}_4\). In the paper, they make the more general conjecture that for fixed \(q\) and \(\gamma\) and all \(g\) sufficiently large, \(N_q(g,\gamma) = \gamma (q+1)\). In a paper that is soon to appear, \textit{F. Vermeulen} has used the theory of curves in toric surfaces to prove this conjecture in the case where \(q\) is odd [``Curves of fixed gonality with many rational points'', Preprint, \url{arXiv:2102.00900}]. \textit{X. Faber} and \textit{F. Vermeulen} have further studied this conjecture in the case where \(q\) is even in [Int. J. Number Theory 18, No. 10, 2211--2216 (2022; Zbl 07572775)]. Finally, in an interesting recent preprint the authors of this paper together with Howe have taken the observation that there does not exist a curve over \(\mathbb{F}_2\) with genus \(5\) and gonality \(6\) as motivation for studying the more general question of when there exists a curve of genus \(g\) and gonality \(g+1\) defined over \(\mathbb{F}_q\).