Vandermonde sets and super-Vandermonde sets (Q958617)

Let \(q=p^e\) be a prime power, and let \(t\) be an integer with \(1 < t < q\). One defines a \(t\)-subset \(\{y_1,y_2,\dots,y_t\}\) of the finite field \(\text{GF}(q)\) to be a \textit{Vandermonde set} if \(\pi_k=\sum_{i=1}^{t}y_i^k = 0\) for all \(1 \leq k \leq t-2\) and a \textit{super-Vandermonde set} if \(\pi_k=0\) for all \(1 \leq k \leq t-1\). Since a \(t\)-subset of a finite field cannot have more than \(t-1\) zero power sums, super-Vandermonde sets are extremal (and only occur when \(p \nmid t\)). In the paper under review it is shown that a super-Vandermonde set of \(\text{GF}(p^2)\) is necessarily a coset of a multiplicative subgroup. In fact, it is shown in general that a super-Vandermonde set of size \(t < p\) or size \(t > q/p\) is necessarily a coset of a multiplicative subgroup. However, as shown by the authors, the classification of all super-Vandermonde sets of size \(t\), for \(p < t < q/p\), would imply the classification of hyperovals, and thus appears to be hopeless at the present time.