Off-diagonal generalized Schur numbers (Q5933472)

The authors state that ``In 1916 Issai Schur proved that any \(r\)-coloring of \([N]=\{1,2,\dots ,N\}\) must admit a monochromatic solution to \(x+y=z\), provided \(N\) is sufficiently large.'' (But \textit{I. Schur} [Über die Kongruenz \(x^m+y^m\equiv z^m\pmod p\). Dtsch. Math. Ver. 25, 114-117 (1916; JFM 46.0193.02)] did not use the language of coloring!) They study numbers they denote by \(S(k_1,k_2,\dots ,k_r)\), defined to be the minimum integer \(M\) such that any \(r\)-coloring of \([M]\) admits, for some \(j\in[r]\), a \(j\)-colored solution to \(\sum_{i=1}^{k_j-1}x_i=x_{k_j}\); the paper is concerned with the numbers \(S(k,\ell)\), which the authors call the off-diagonal generalized Schur numbers or Issai numbers [\textit{A. Robertson}, Adv. Appl. Math. 25, No. 2, 153-162 (2000; Zbl 0957.05076)]. They attribute the \(S\) notation with \(r\) arguments, and the term generalized Schur numbers to \textit{A. Beutelspacher} and \textit{W. Brestovansky} [Lect. Notes Math. 969, 30-38 (1982; Zbl 0498.05002)] when, in fact, those authors called their numbers ordinary Schur numbers, reserved the designation generalized for an extension to arithmetic progressions, and used a different notation. It is observed that Beutelspacher and Brestovansky proved that \(S(k,k)=k^2-k-1\) \((k\geq 3)\). ``We show that \(S(3,\ell)=3\ell-4\) for odd \(\ell\geq 3\), \(S(3,\ell)=3\ell-5\) for even \(\ell\geq 4\), and \(S(k,\ell)=k\ell-\ell-1\) for \(4\leq k\leq\ell\)''.