The major requirements for pure fusion tokamak reactors and tokamak-based fusion neutron sources (FNS) are analyzed together with possible paths from the present-day tokamak towards the FNS tokamak. The FNS are of interest for traditional fission reactors as a method of waste management by burning of long-lived transuranic radionuclides (minorities) and fission fuel breeding. The Russian fission community places several hard requirements on the quality of FNS suitable for the first step of the investigation program of minority burning and breeding. They are (a) a steady-state regime of neutron production (more than 80% of the operational time), (b) a neutron power flux density greater than > 0.2 MWm(-2), (c) a total surface integrated neutron power > 10MW. Among the different FNS projects, based on magnetically confined plasmas, only 'classical tokamak' is most likely to fulfill these requirements in the nearest future. Some of the most important improvements of the 'classical tokamak' needed for successful realization of the FNS are (1) decrease in Z(eff) (probably, by making use of lithium as a part of plasma-facing components), (2) He removal and closed loop DT fuel circulation, (3) increase in the energy of stationary injected neutral tritium beams up to 150-170 keV and (4) control of impurity contamination at the plasma center (probably, by local RF heating). These key issues are discussed.