Browse Prior Art Database

Automatic Programming

IP.com Disclosure Number: IPCOM000076487D
Original Publication Date: 1972-Mar-01
Included in the Prior Art Database: 2005-Feb-24
Document File: 4 page(s) / 88K

Publishing Venue

IBM

Related People

Woodland, NJ: AUTHOR

Abstract

A program system is described which will enable a digital computer to discover the mathematical formulations underlying application example data, and then to automatically prepare a computer program for the application. The objects of this system may be accomplished in general by any stored program digital computer operating under an object Program coded according to the flow chart of Fig. 1, which includes the six subroutines flow-charted in Figs. 2-7.

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 54% of the total text.

Page 1 of 4

Automatic Programming

A program system is described which will enable a digital computer to discover the mathematical formulations underlying application example data, and then to automatically prepare a computer program for the application. The objects of this system may be accomplished in general by any stored program digital computer operating under an object Program coded according to the flow chart of Fig. 1, which includes the six subroutines flow-charted in Figs. 2-7.

The flow chart of Fig. 1 is for a program which integrates a number of subroutines into an embodiment which enables a computer to analyze numeric examples of inputs and outputs of a data processing task; to identify the mathematical functions implicit in the data; to describe branching logic implicit in the data; and to identify irregularities in the data.

The subroutines of Figs. 2, 3, 5, 6 and 7 will be readily understood by those skilled in the art of programming.

The SEARCH subroutine of Fig. 2 enables the computer to arrange examples data in a manner which will facilitate the analytical operations of function identification as set forth in Fig. 4.

The MOVE subroutine of Fig. 3 enables the computer to bypass examples.

The SIEVE subroutine of Fig. 4 enables the computer to identify functions implicit in input and output data of numeric examples of a data processing task. SIEVE is called by the subroutines of Figs. 2 and 7. Each of those calling subroutines provides n examples as input to subroutine SIEVE. The value n must be equal to or greater than the minimum number of examples, or observations, which is required to enable regression analysis of the examples data; it is a function of the number of parameters involved and the method of computing its minimum value is well known to those skilled in statistics. In the coding of Fig. 10, it is assumed that the number of independent parameters does not exceed three. In this case n=5 is great enough and five is the number of examples in Fig. 8.

The BRANCH subroutine of Fig. 5 enables the computer to identify a pa...