Approximation results in the strict topology (Q1381935)

For a topological space \(X\), \(B_0(X)\) denotes the set of all nonnegative bounded functions on \(X\) which vanish at infinity. The strict topology \(\beta\) on the space \(C_b(X, E)\) of all bounded continuous functions on \(X\) with values in the real normed space \(E\) is given by the family \((p_\varphi)_{\varphi\in B_0(X)}\) of seminorms, where \(p_\varphi(f)= \sup_{x\in X}\varphi(x)\| f(x)\|\); here one can clearly restrict attention to \(\varphi\in D_0(X)= \{f\in B_0(X); 0\leq f\leq 1\}\). In the sequel, let \(W\) be a nonempty subset of \(C_b(X, E)\). A continuous function \(\psi\) on \(X\) with \(0\leq\psi\leq 1\) is called a multiplier of \(W\) if \(\psi f+ (1-\psi)g\in W\) for all \(f,g\in W\). \(W\) is said to be \(\beta\)-localizable under a subset \(S\) of \(C_b(X,\mathbb{R})\) if for every \(f\in C_b(X, E)\) the following two conditions (i) and (ii) are equivalent: (i) \(f\) belongs to the \(\beta\)-closure of \(W\), (ii) for every \(\varphi\in D_0(X)\), every \(\varepsilon> 0\) and every \(x\in X\), there is \(g_x\in W\) such that \(\varphi(t)\| f(t)- g_x(t)\|< \varepsilon\) for all \(t\in [x]_S= \{y\in X\); \(h(x)= h(y)\) for each \(h\in S\}\). It is known (for locally compact \(X\) see \textit{J. B. Prolla} [``Approximation of vector valued functions'', North-Holland Math. Stud. 25 (1977; Zbl 0373.46048)]) that for a completely regular Hausdorff space \(X\) a vector subspace \(W\) of \(C_b(X, E)\) which is an \(A\)-module over a subalgebra \(A\) of \(C_b(X, \mathbb{R})\) is \(\beta\)-localizable under \(A\). Now, the authors reprove several known approximation theorems for the strict topology and with similar methods derive results of the following type: \(W\) is \(\beta\)-localizable under the set \(A\) of all multipliers of \(W\) if (a) \(W\) is a uniformly bounded subset of \(C_b(X, E)\) or (b) \(W\) is a nonempty convex subset of \(C_b(X, E)\), \(X\) a completely regular Hausdorff space. For compact \(X\) the results can already be found in \textit{J. B. Prolla} [Weierstrass-Stone, the theorem, Frankfurt/Main (1993; Zbl 0849.41029)], and the present proofs follow rather directly from this case. There are many obvious misprints.