In this report we discuss the merits and demerits of the formal Hamiltonian approach for spherically symmetric thin shells and methods of quantization connected with it. The merits are: (1) we are using the approved way of quantization, which algorithm was filed earlier; (2) we do not know other ways. The demerits are: (1) we attempt to introduce a time into the of essence stationary theory, that leads to the “many times” and “many spectra” problems, and also to the “where is observer” question; (2) obtained wave functions are not globally defined, that leads to the “what about boundary conditions” question. So, we propose a way of the direct quantization, which consists not in the quantization by means of artificial Hamiltonians constructing but in the immediate quantization of a conservation law. Herewith we will have: (1) all values are observable and have an obvious physical meaning; (2) wave function is globally defined and often can be exactly obtained; (3) mathematically we expedite the problem and, therefore, more complex tasks can be considered exactly. 04.20.-q, 04.60.+n, 11.17.+y Typeset using REVTEX Present electronic address: email@example.com (Sub: for Zlo), all others are temporary.