Carlo John Evangelisti

System partitioning and its measure

Program modules and data structures are interconnected by calls and references in software systems. Partitioning these entities into clusters reduces complexity. For very large systems manual clustering is impractical. A method to perform automatic clustering is described and a metric to quantify the complexity of the resulting partition is developed.

GREENPRINT: a graphic representation of structured programs

To improve the readability of programs over existing techniques, a new program representation termed GREENPRINT has been developed and is discussed in this paper. GREENPRINTs (the name taken from the phosphor fluorescence of certain display terminals and paralleling the term blueprints) are tree-structured diagrams together with source code statements that represent the control structure of programs. Discussed in this paper are the diagramming conventions, control flow methodology, presentation graphics, and practical experience with GREENPRINTs.

A computer graphics system for block diagram problems

An experimental on-line network design system is proposed. Called DESIGNPAD, it consists of a small computer with graphic display equipment connected to a time-sharing computer and includes the necessary programming support for the equipment. The system is designed to accept problems covering a broad spectrum of applications in the form of labeled block diagrams. The input/output medium, the man-machine interface, and the supporting data structures, particularly the cellular structure, are discussed.

Some experiments in the evaluation of a character recognition scanner

Abstract For character recognition machines in which the design of the logic is time consuming and expensive, a method for evaluating the character recognition scanner prior to designing the recognition logic is desirable. A method for evaluating a character recognition scanner by evaluating the patterns it produces is described. The method, for a given set of parameters, assigns to each pattern a figure of merit that measures the difficulty of recognizing the pattern. A number of experiments show that the evaluation method produced results comparable to results produced by a recognition logic designed by engineers. On a sample of 81,200 patterns, the evaluation method showed in this experiment that 0.20% of the patterns would be difficult to recognize, while for the same patterns, logic designed by engineers failed to recognize or misrecognized 0.24% of the patterns.

Some characteristics of a TDI scanner

We develop methods to characterize the noise positional accuracy and color shift for a camera type scanner. These methods have been applied to several similar scanners based on an experimental Time Delay Integration chip developed at IBM Research. The scanner is further discussed in the paper by Hon-Sum Wong et al in this session. To measure noise we scan a field twice under identical conditions. We then consider the noise to be the difference in these measurements computed pixel by pixel. All data is measured with a 12 bit a/d converter (0-4095). For a typical dark current correction we find a standard deviation of 1. 97. For a white field we find the standard deviation of the measurements to be 3. 65 out of the total range of 4095 (this is . 089). To measure color shift and positional accuracy we scan a transmissive test image which consists of an aluminized glass plate in which 6 rows of 6 diamonds per row all on a square grid were engraved. The position of a diamond may be accurately obtained for each color scan by summing the gray values in the various rows and in the various columns of an area containing one diamond. We find color shifts of up to . 5 pixel. The position accuracy is good to . 2 pixel if the slight rotation of the TDI device relative to