How does the newly proposed DATABASE_PDB_MATRIX handling of SCALE differ from _atom_sites.fract_transf_matrix ... and _vector ...? Also, would it not be clearer to put the ORIGX information in as a sub-category of atom_sites, say, as .PDB etc., etc., since in fact it is primarily an atom_sites issue and only secondarily (if at all) a database issue. It is a transform to bring the dataset back to a coordinate frame the experimenter thought useful. -- H. J. Bernstein