Joe W. Duran

An Evaluation of Random Testing

Random testing of programs has usually (but not always) been viewed as a worst case of program testing. Testing strategies that take into account the program structure are generally preferred. Path testing is an often proposed ideal for structural testing. Path testing is treated here as an instance of partition testing, where by partition testing is meant any testing scheme which forces execution of at least one test case from each subset of a partition of the input domain. Simulation results are presented which suggest that random testing may often be more cost effective than partition testing schemes. Also, results of actual random testing experiments are presented which confirm the viability of random testing as a useful validation tool.

Capture-Recapture Sampling for Estimating Software Error Content

Mills capture-recapture sampling method allows the estimation of the number of errors in a program by randomly inserting known errors and then testing the program for both inserted and indigenous errors. This correspondence shows how correct confidence limits and maximum likelihood estimates can be obtained from the test results. Both fixed sample size testing and sequential testing are considered.

On some reliability estimation problems in random and partition testing

Random testing is receiving increasing attention in recent years. Aside from its relative simplicity and low cost, studies have shown that random testing is an effective testing strategy. An advantage of random testing is that the reliability of the program can be estimated from the test outcomes. The authors extend the Thayer-Lipow-Nelson reliability model to account for the cost of errors. They also compare random with partition testing by looking at upper confidence bounds for the cost weighted performance of the two strategies.<<ETX>>

A system for automatic acquisition of three-dimensional data

This paper presents the design of a three-dimensional data acquisition system based on multiple, single-dimensional optical sensors. The system can operate in any of three modes: (1) the tracking of multiple, independent, point light sources (2) the automatic digitization of opaque surfaces (3) the real-time tracking of an unmarked moving object (e.g., tip of users hand). The design offers such advantages as a lensless sensing system, a minimum reliance on analog measurements, an ease of upgrading to higher precision measurements, an ease of portability, an adjustable field of view, and the ability to operate under normal ambient light conditions. A network of microprocessors is incorporated to minimize processing delays and thus increase data acquisition rates. In its initial application the system will digitize the cranio-facial surfaces of candidates for reconstructive surgery.

Toward models for probabilistic program correctness

Program testing remains the major way in which program designers convince themselves of the validity of their programs. Software reliability measures based on hardware reliability concepts have been proposed, but adequate models of software reliability have not yet been developed. Investigators have recently studied formal program testing concepts, with promising results, but have not seriously considered quantitative measures of the “degree of correctness” of a program. We present models for determining, via testing, such probabilistic measures of program correctness as the probability that a program will run correctly on randomly chosen input data, confidence intervals on the number of errors remaining in a program, and the probability that the program has been completely tested. We also introduce a procedure for enhancing correctness estimates by quantifying the error reducing performance of the methods used to develop and debug a program.