Waring's problem for polynomial rings and the digit sum of exponents (Q515994)

The paper contains some results on a classical variant of Waring's problem. Namely on the so-called \textit{easy} Waring's problem for polynomials. More precisely, let \(p\) be a prime number, \(q\) be a power of \(p\), say \(q := p^n\), let \(\mathbb{F}\) be a fixed algebraic closure of the finite field \(\mathbb{F}_q\) and let \(k \in \mathbb{N}\) be non divisible by \(p.\) Moreover, set \(g(k,q)\) the sum of all digits of \(k\) written in basis \(q.\) Assume that \(t \in \mathbb{F}[t]\) is a sum of \(k\)-th powers of polynomials in \(\mathbb{F}[t].\) Put \(v(k,q)\) the least positive integer \(m\) such that \(t\) is a sum of \(m\) \(k-\)th powers of polynomials in \(\mathbb{F}[t].\) Two conditions are required (a) \(g(k,q) \leq M\) for some \(M \in \mathbb{N}\) and (b) \(M-1\) divides \(q-1.\) In this case one has (i) \(v(p,k) \leq 2M-1\) when \(g(k,q)=M\) and (ii) \(v(p,k) \leq 2M-2\) when \(g(k,q)< M.\) The long standing related open problems : (a) Is \(t\) a sum of three cubes of polynomials with coefficients in the finite field of sixteen elements and (b) Is \(t\) a sum of three cubes of polynomials with rational coefficients are not mentioned.