On the nonexistence of ternary linear codes attaining the Griesmer bound (Q2119947)

Given the finite space \(\mathbb{F}_q^n\) for prime power \(q\), the Griesmer bound \(n\geq g_q(k,d)=\sum\limits_{i=0}^{k-1}\lceil d/q^i\rceil\) is attained for all sufficiently large \(d\) for fixed \(q\) and \(k.\) The authors are tacking the problem to find \(D_{q,k}\), the largest value of \(d\) such that the Griesmer bound is not attained for fixed \(q\) and \(k.\) This problem is already solved for the cases \(q \geq k\) with \(k = 3, 4,5\) and \(q\geq 2k-3\) with \(k\geq 6,\) but not known for the case \(q < k\) except for some small \(q\) and \(k.\) In this paper it is shown that the conjecture \(D_{3,k}=T_k\) and \(n_3(k, T_k)=g_3(k,T_k)+1\) for all \(k \geq 4\) is valid for \(k\leq 9.\) This is achieved by showing that: (A) There exists no \([g_3(k, T_k ), k, T_k]_3\) code; (B) There exists a \([g_3(k, T_k ) + 1, k, T_k]_3\) code. Geometric methods to construct codes or to prove the nonexistence of codes with certain parameters are used. Particularly, it is proved that (B) is valid for all \(k\geq 4\) and that (A) is valid for \(k=8,9.\)