Several trends have recently combined to escalate the importance of quantitative software reliability measures: the large and growing number of real-time and interactive systems has incresed the operational and cost i...
详细信息
Several trends have recently combined to escalate the importance of quantitative software reliability measures: the large and growing number of real-time and interactive systems has incresed the operational and cost impacts of failures; the increasing number, size, and complexity of computer networks and distributed processing systems have multiplied the risk effects of failure; the explosive growth of personal computing has created a demand for relatively foolproof software for unsophisticated users. The theory outlined in this paper has proved to be a good framework for understanding, measuring, and predicting the reliability of computer programs. Improvement of reliability usually impacts development schedules and system performance (in the case of software, through increased memory, processing time, and peripherals requirements). The system engineer and the manager have to make design tradeoffs, and it is best that this be done in quantitative terms. The need for a quantitative reliability measure continues throughout the development process, particularly during test, since reliability is a valuable indicator of system status. Reliability or mean-time-to-failure (MTTF) is a useful metric for characterizing system operation and for controlling change during the maintenance phase. This paper focuses on the system engineering application, but it also touches on monitoring the system test phase and controlling change during maintenance.
Concise and natural verification systems are developed for handling control structures designed to aid in structuring programs. Also, an interpretative semantics is developed for justifying the axioms. Two principles ...
详细信息
Concise and natural verification systems are developed for handling control structures designed to aid in structuring programs. Also, an interpretative semantics is developed for justifying the axioms. Two principles are exploited in the development of axioms for block exit control structures. The first is the information (in the form of assertions) must be maintained for both sequential and nonsequential (i. e. , exits) executive sequences. The second principle is that destination information (for exits) must be used in order to select the correct assertion as precondition for the statement following the block exited. Since all the multiple exit blocks are labeled, labels are used to perform this selection.
Over the past six years, research and development have primarily centered on the problem of converting in non-dynamic environments. A survey is made of the tools and technology currently available for the conversion o...
详细信息
Over the past six years, research and development have primarily centered on the problem of converting in non-dynamic environments. A survey is made of the tools and technology currently available for the conversion of data organization. Database program conversion is reviewed from the standpoint of the research directions being taken by current research. The factors affecting conversion, the approaches for reducing the need to convert data and applications programs, and the impact of new software and hardware technologies on conversion are discussed. The trends in conversion needs and too are summarized.
During the period Aug 1974 to May 1978 a study to evaluate the accuracy of predictions of several models for predicting the error content and reliability of a software package against error data extracted from four la...
详细信息
During the period Aug 1974 to May 1978 a study to evaluate the accuracy of predictions of several models for predicting the error content and reliability of a software package against error data extracted from four large software development projects was undertaken. This paper briefly describes the results of this empirical study for three such models, the Jelinski-Moranda, Schick-Wolverton and a modified Schick-Wolverton. Model predictions are analyzed and general conclusions drawn as to model applicability.
The estimation process of software reliability is described and its various phases are discussed. Three models are suggested for the error detection process. These include the Instant Correction, the Delayed Correctio...
详细信息
The estimation process of software reliability is described and its various phases are discussed. Three models are suggested for the error detection process. These include the Instant Correction, the Delayed Correction, and the Random Correction models. The estimators for these models are described along with some experimental results. Also discussed are the problems of confidence intervals and validity testing.
Two compatible algorithms for on-line division and multiplication, based on the redundant number systems, are presented. Besides an obvious usefulness in real-time applications, these on-line algorithms provide an eff...
详细信息
Two compatible algorithms for on-line division and multiplication, based on the redundant number systems, are presented. Besides an obvious usefulness in real-time applications, these on-line algorithms provide an effective way of speeding up the execution of sequences of the basic arithmetic operations by minimizing the delay between successive operations in an overlapped mode of operation. The algorithms can be seen to have rather simple implementation requirements and properties which are compatible with the desirable modularity in implementation and variable precision operations.
The benchmark is a vital part of the competitive selection process. Here a Benchmark Preparation System is described to reduce the nonportability and expense of using natural benchmarks without losing the characterist...
详细信息
The benchmark is a vital part of the competitive selection process. Here a Benchmark Preparation System is described to reduce the nonportability and expense of using natural benchmarks without losing the characteristics of the users workload in terms of processing efficiency and representation. The results obtained indicate that these objectives can be met. Based on efforts, it is believed that portability can be achieved by an automated means without sacrificing the efficiency of a computer system.
An algorithm is presented for constructing a tree to satisfy a set of lineage constraints on common ancestors. This algorithm is then applied to synthesize a relational algebra expression from a simple tableau, a prob...
详细信息
An algorithm is presented for constructing a tree to satisfy a set of lineage constraints on common ancestors. This algorithm is then applied to synthesize a relational algebra expression from a simple tableau, a problem arising from the theory of relational databases.
暂无评论