On mappings preserving inner product modulo an ideal (Q1125390)

The authors investigate the functional equation of inner product preserving mappings, that is the equation \[ \langle f(x) |f(z) \rangle_Y = \langle x|z \rangle_X \qquad (x,z)\in X^2\setminus M, \] where \(X\), \(Y\) are inner product spaces, \(f:X\to Y\), and \(M\) is a ``small'' subset of \(X^2\) in a certain sense. If \(M\) is the empty set, then \(f\) is called inner product preserving. In a former paper \textit{J. Chmieliński} and \textit{J. Rätz} [Publ. Math. Debrecen 52, No. 3-4, 317-335 (1998; Zbl 0909.39010)] it was shown that under properly defined smallness in \(X^2\) and \(X\), the almost inner product preserving mappings are almost everywhere equal to a (uniquely determined) inner product preserving mapping. In the present paper, using a different technique, results of the same kind are obtained, however the notion of smallness used here is much more general. The authors can eliminate the following very restrictive assumptions from the notion of smallness of subsets of \(X^2\): {(a)} almost all sections in \textit{each} direction (not only vertical) are assumed to be ``small'' in \(X\); {(b)} two individual sections along the main diagonal and (in case \(\mathbf K=\mathbb C\)) along the line \(\{(ix,x):x\in X\}\) are supposed to be ``small''. Thus the results obtained essentially generalize that of the paper mentioned above.