Mappings of type Orlicz and generalized Cesáro sequence space (Q394636)

Let \(E\) be a solid linear space of real sequences \(x=(\xi_0,\xi_1,\xi_2,\dots )\) that is invariant under the operator \( (\xi_0,\xi_1,\xi_2,\dots )\mapsto(\xi_0,\xi_0,\xi_1,\xi_1,\xi_2,\xi_2,\dots )\). Using the approximation numbers \(a_n(T)\) of operators \(T\) from a Banach space \(X\) into a Banach space \(Y\), the authors prove in Theorem~3.1 that the components \(U_E^{\mathrm{app}} (X,Y) := \big\{ T \in L (X,Y) : \big(a_n(T)\big) \in E \big\}\) form an operator ideal. Since one needs only the cone \(C := \big\{ x \in E : \xi_0 \geq \xi_1\geq \xi_2\geq \dots \geq 0 \big\}\), this result has been essentially known for more than 50 years. In the case of a Hilbert space, it can be found on p.~26 of [\textit{R. Schatten}, Norm ideals of completely continuous operators. Berlin-Göttingen-Heidelberg: Springer-Verlag (1960; Zbl 0090.09402)]. The extension to the setting of Banach spaces is a triviality. In what follows, we mention only a few monstrosities contained in this paper. As regards Orlicz spaces, the authors had bad luck with the \(\Delta_2\)-condition. When they assumed this condition (Theorem~3.2 and Cor.~3.8), it is superfluous. When they didn't assume this condition (Theorem~3.4), it is needed. Concerning Orlicz sequence ideals and the associated operator ideals, we refer to pp.~189--191 of [\textit{A. Pietsch}, Operator Ideals. North-Holland Mathematical Library 20 (1980; Zbl 0434.47030)]. Theorem 3.7 says that \(U_E^{\mathrm{app}} (X,Y)\) is complete for every pre-modular special sequence space \(E\). Strangely enough, the definition of those spaces \(E\) does NOT include completeness. Hence it is clear that the corresponding space of operators need not be complete. Comparing the exponents in lines~2 and 3 of the equation array in step~(3) of the ``proof'' of Theorem~3.3 shows that even elementary manipulations are false. The mysterious condition \(\lim_{n\to \infty} \sup p_n < \infty\), which occurs in the formulation of Theorem~3.3, seems to mean \(\limsup_{n\to\infty} p_n < \infty\). However, throughout the paper, \((p_n)\) is assumed to be a bounded sequence of positive numbers. Finally, the reviewer cannot resist quoting the last sentence of the paper: ``The authors wish to thank the referees for their careful reading of the paper and for their helpful suggestions.''