Some more recent results concerning weak Asplund spaces (Q2495007)

It is shown (in ZFC) that there exists a Gâteaux differentiability space which is not weak Asplund. A Banach space is a weak Asplund space, respectively a Gâteaux differentiability space, if each convex function defined on it is Gâteaux differentiable on a residual, respectively dense, subset. The proof uses the notions of Stegall and weakly Stegall topological spaces (whose definitions involve the notion of USCO mappings defined on Baire spaces); a Banach space \(X\) is weak Asplund whenever \((X^\ast, \text{weak}^\ast)\) is a Stegall space and is a Gâteaux differentiability space (and even an almost weak Asplund space) whenever \((X^\ast, \text{weak}^\ast)\) is a weak Stegall space. The key of the proof is the construction of some specific everywhere second category subset \(A\) of \((0,1)\). Then, if \(K_A\) is the Kalenda compact associated to \(A\) (\(K_A\) is a generalization of the ``double arrow'' space, which corresponds to the case \(A = (0,1)\)), the space \(X = {\mathcal C}(K_A)\) is a Gâteaux differentiability space (and even \(({\mathcal C}(K_A)^\ast, \text{weak}^\ast)\) is weakly Stegall), but \(X\) is not weak Asplund.