Menu organization through block clustering

Abstract

Block Clustering is proposed as a method to derive the semantic relationships among entities that comprise application programs, operating system environments or programming environments. The procedure was derived from the literature on clustering in the biological sciences. The advantages of the procedure are that it relies on information obtained from design specifications, eliminating difficulties associated with designing through “expert” opinion, and allowing the procedure to be implemented early in the design of a system. Also the procedure works well with systems possessing a small or a large number of entities. Two cases studies of the Block Clustering method are presented for two different fourth generation programming languages. Specificially the 37 HyperTalk 1.5 properties were block clustered and found to have six menu topics. Also the 98 SuperTalk 1.5 properties were Block Clustered and found to have 11 menu topics. The results illustrate the utility of Block Clustering analysis for both small and large systems.