We have been developing a multi-purpose pattern-fitting system RIETAN-FP and a three-dimensional visualization system VENUS, which have been extensively used for structure refinements of various metal and inorganic materials from neutron powder diffraction data. At first, their outlines and the history of their developments are shortly looked back. The second part describes procedures for analyzing collinear magnetic structures with the combination of VESTA in the VENUS system and RIETAN-FP by taking BiCoO3 for instance. Finally, a new C++ program, Dysnomia, for the maximum entropy method is introduced with emphasis on its new features. Dysnomia excels its predecessor, PRIMA, in computation speed, memory efficiency, scalability, and reliability. In particular, addition of a normal-distribution constraint is effective in obtaining nuclear-density distribution that is physically and chemically reasonable.