“…In the case of "very" discontinuous initial data (e.g., the kernel K(s,t) is only Σ-measurable in s, and/or y G L°°(D)) we are forced to look for a solution in the space 354 R. Lepp of bounded measurable functions (see, e.g., [15]). Then the natural question arises: how to understand the convergence of solutions of approximate problems to the solution of the initial problem, since the well-known projection methods do not work in L°°(D) (the space C(D) of continuous functions is not dense in L°°(D) and, thus, it is not clear how to construct a system of "nice" subspases {B n }, B\ C £ 2 C .…”