Logarithmic transformations into \(l^2\) (Q1769334)

In 1958 D. Borwein introduced the logarithmic power series method of summability. Its matrix analogue, denoted by \(L_t\) has been recently considered by the author [Rocky Mt. J. Math. 28, No. 1, 253--266 (1998; Zbl 0922.40007)], by \((L_t(u))_n= -{1\over\log(1- t_n)}\,\sum^{+\infty}_{k=0} {1\over k+1} u_k t^{k+1}_n\), where \(0< t_n< 1\) for all \(n\), and \(\lim_{n\to\infty}\,t_n= 1\) (here \((L_tu)_n\) is the \(n\)th term of the sequence \(L_t u\)). In the present paper, the author studies \(L_t\) as mappings into \(\ell^2\), where \(\ell^2= \{u: \sum^{+\infty}_{k=0} |u_k|^2<+ \infty\}\). Necessary and sufficient conditions for \(L_t\) to be \(\ell^2-\ell^2\) (i.e. \(L_tu\) is in \(\ell^2\) whenever \(u\) is in \(\ell^2\)), as well as similar results, are proved.