Are complete K-trees something more than decision trees?
Gabriel Valiente · ACM SIGART Bulletin · 1991
I have just read the research report "Using Complete K-Trees to Generate Code in Pascal for an Expert System" by Dr. Edward N. Schwartz, appeared in SIGART Bulletin 1(2): 12--13. The report first presents K-trees as an "implementation independent method of knowledge representation," useful "in all stages of expert system development." The idea of a complete K-tree is then introduced, in which all possible values of every property are explicitly recorded in the tree. As this is the central result presented in the report, I think it deserves some comments.