Skip to main content

Amiram Yehudai

Followers

9

Following

6

Co-authors

6

Public Views

Shmuel Tyszberowicz

Daniel M Zimmerman

University College London

Laurent Fournier

IBM Research

Benny Pasternak

Tel Aviv University

Tomohiro Otsuki

Colorado State University

anuranjan misra

Uttar Pradesh Technical University

Interests

Uploads

Papers by Amiram Yehudai

OBSERV/CRUISE - A CASE environment for prototyping of complex systems

by Amiram Yehudai and Shmuel S Tyszberowicz

. The paper presents a CASE environment for rapid prototyping of software. The environment suppor... more . The paper presents a CASE environment for rapid prototyping of software. The environment supports the OBSERVmethodologyand language,which combines the objectoriented approach (for system decomposition), finite state machines (to model behavior), and logic programming paradigm (for describing activities). The environment is based on a graphical front end, named CRUISE and includes tools for editing, simulation (or execution), browsing, and static checking. The paper outlines the OBSERV methodolgy, and presents in detail the CRUISE user interface, and each of the tools. 1 Introduction The traditional software development process is unsuitable for complex target systems (e.g. embedded systems). Rapid prototyping is increasingly seen as an important software development paradigm, augmenting or even replacing the conventional phased approach [3, 5, 13]. The OBSERV methodology for software development is based on rapid construction of an executable specification (or prototype) of ...

Locating Regression Bugs

by Shmuel S Tyszberowicz and Amiram Yehudai

Lecture Notes in Computer Science, 2008

ABSTRACT A regression bug is a bug which causes a feature that worked correctly to stop working a... more ABSTRACT A regression bug is a bug which causes a feature that worked correctly to stop working after a certain event (system upgrade, system patching, daylight saving time switch, etc.). Very often an encompassed bug fix included in a patch causes the regression bug. Regression bugs are an annoying and painful phenomena in the software development process, requiring a great deal of effort to find. Many tools have been developed in order to find the existence of such bugs. However, a great deal of manual work is still needed to find the exact source-code location that caused a regression bug. In this paper we present the CodePsychologist, a tool which assists the programmer to locate source code segments that caused a given regression bug. The CodePsychologist goes beyond current tools, that identify all the lines of code that changed since the feature in question worked properly (with the help of a Source Control Tool). The CodePsychologist uses various heuristics to select the lines most likely to be the cause of the error, from these often large number of lines of code. This reduces the fixing time of regression bugs. It allows a quick response to field errors that need an immediate correction.

KeYGenU: combining verification-based and capture and replay techniques for regression unit testing

by Amiram Yehudai and Shmuel S Tyszberowicz

International Journal of System Assurance Engineering and Management, 2011

Unit testing plays a major role in the software development process. Two essential criteria to ac... more Unit testing plays a major role in the software development process. Two essential criteria to achieve effective unit testing are: (1) testing each unit in isolation from other parts of the program and (2) achieving high code coverage. The former requires a lot of extra work such as writing drivers and stubs, whereas the latter is difficult to achieve when manually writing the tests. When changing existing code it is advocated to run the unit tests to avoid regression bugs. However, in many cases legacy software has no unit tests. Writing those tests from scratch is a hard and tedious process, which might not be cost-effective. This paper presents a tool chain approach that combines verification-based testing (VBT) and capture and replay (CaR) test generation methods. We have built a concrete tool chain, KeYGenU, which consists of two existing tools -KeY and GenUTest. The KeY sys-tem is a deductive verification and test-generation tool. GenUTest automatically generates JUnit tests for a correctly working software. This combination provides isolated unit test suites with high code-coverage. The generated tests can also be used for regression testing.

Generating Regression Unit Tests Using a Combination of Verification and Capture & Replay

by Shmuel S Tyszberowicz and Amiram Yehudai

Lecture Notes in Computer Science, 2010

The combination of software verification and testing techniques is increasingly encouraged due to... more The combination of software verification and testing techniques is increasingly encouraged due to their complementary strengths. Some verification tools have extensions for test case generation. These tests are strong at detecting software faults during the implementation and verification phase, and to further increase the confidence in the final software product. However, tests generated using verification technology alone may lack some of the benefits obtained by using more traditional testing techniques. One such technique is Capture and Replay, whose strengths are the generation of isolated unit tests and regression test oracles. Hence, the two groups of techniques have complementary strengths, and therefore are ideal candidates for a tool-chain approach proposed in this paper. The first phase produces, for a given system, unit tests with high coverage. However, when using them to test a unit, its environment is tested as well -resulting in a high cost of testing. To solve this problem, the second phase captures the various executions of the program, which are monitored by the output of the first phase. The output of the second phase is a set of unit tests with high code coverage, which uses mock objects to test the units. Another advantage of this approach is the fact that the generated tests can also be used for regression testing.

KeYGenU: combining verification-based and capture and replay techniques for regression unit testing

by Shmuel Tyszberowicz, Bernhard Beckert, and Amiram Yehudai

International Journal of System Assurance Engineering and Management, 2011

Unit testing plays a major role in the software development process. Two essential criteria to ac... more Unit testing plays a major role in the software development process. Two essential criteria to achieve effective unit testing are: (1) testing each unit in isolation from other parts of the program and (2) achieving high code coverage. The former requires a lot of extra work such as writing drivers and stubs, whereas the latter is difficult to achieve when manually writing the tests. When changing existing code it is advocated to run the unit tests to avoid regression bugs. However, in many cases legacy software has no unit tests. Writing those tests from scratch is a hard and tedious process, which might not be cost-effective. This paper presents a tool chain approach that combines verification-based testing (VBT) and capture and replay (CaR) test generation methods. We have built a concrete tool chain, KeYGenU, which consists of two existing tools -KeY and GenUTest. The KeY sys-tem is a deductive verification and test-generation tool. GenUTest automatically generates JUnit tests for a correctly working software. This combination provides isolated unit test suites with high code-coverage. The generated tests can also be used for regression testing.

Combining reflection and finite state diagrams for design enforcement

by Vered Gafni and Amiram Yehudai

Theory and Practice of Object Systems, 1996

This paper presents a method for design enforcement, based on a combination of re ection and Stat... more This paper presents a method for design enforcement, based on a combination of re ection and State Machine Diagrams. The method is supported by RECOS, a RE ective system in a Concurrent Object-oriented environment that uses Finite State Diagrams. The re ection methodology is now a developing branch in computer science for controlling and modifying systems, particularly Object-Oriented systems. RECOS is a \base-system" under which other systems can be built. Combining concepts of Concurrent Object-Oriented design, Finite State Diagrams, and re ection leads to an increase in the reliability of the systems, by insuring that objects function according to their design. RECOS also presents a safe way for handling modi cations of the systems. We attempt to provide the developer of a system under RECOS with an environment that balances between easing implementation e ort and guarding against unsafe implementations. We address here the issue of re ection, present the structure of RECOS, using as a running example a heating system, and brie y discuss implementation details and future work.

Reasoning About Real-Time Reactive Systems with MASS

by Yishai Feldman, Amiram Yehudai, and Vered Gafni

This paper lays the theoretical foundation for formal reasoning with the language MASS. MASS is a... more This paper lays the theoretical foundation for formal reasoning with the language MASS. MASS is an activation oriented speci cation language, suitable for specifying reactive real time systems. Following a short review of MASS, the paper presents a logical formalism based on the language of MASS events, for stating properties of MASS speci cations. A version of MASS in which speci ed reactions must be given explicit deadlines is then shown to be decidable. The general unbounded version is shown to be undecidable.

Formal Development of Real-time Systems with PLOT/MASS

by Yishai Feldman, Amiram Yehudai, and Vered Gafni

The paper presents a formal development process of a real-time system. We employ two related lang... more The paper presents a formal development process of a real-time system. We employ two related languages: the real-time logic plot for plant modeling and requirements speci cation, and the real-time language mass for the controller design. A deductive proof system is used to verify the design against the requirements. The examples given in the paper demonstrate the special features of the mass/plot framework that make it especially suitable for real-time systems speci cation. In particular, we show explicit expression of causality and timing in plot, modular design by re nement in mass, and the succinct and intuitive nature of speci cations in this framework.

Real-Time Systems Development with MASS??

by Yishai Feldman, Amiram Yehudai, and Vered Gafni

Electronic Notes in Theoretical Computer Science, 1999

In this paper, we demonstrate the capability of mass, a real-time design language, for large syst... more In this paper, we demonstrate the capability of mass, a real-time design language, for large systems speci cation. The paper presents a hierarchical speci cation of an automatic cruise controller that evolves through stepwise re nement. In particular, we show modular design, the separation of the functional and reactive concerns, and the succinct and intuitive nature of speci cations in mass.

Activation-oriented specification of real-time systems

by Yishai Feldman, Amiram Yehudai, and Vered Gafni

Lecture Notes in Computer Science, 1994

We p r e s e n t MASS, a declarative language for specifying the reactive b e h a vior of real-ti... more We p r e s e n t MASS, a declarative language for specifying the reactive b e h a vior of real-time systems. The basic primitive of the language is the task, which i s t h e i n terface between the reactive and sequential aspects of the speci cation. The purely computational meaning of a task (as an I/O transformation) is given outside MASS using standard speci cation languages for sequential computations. The reactive aspects of real-time systems are expressed in MASS through causal and temporal relations between events that signal task terminations. Hierarchical systems are obtained by re ning tasks, specifying them as sub-systems in MASS. MASS is given a formal semantics in a trace model augmented with explicit representation of causes. A synchronous execution model conforming with the formal semantics makes MASS speci cations executable. A development system based on this model allows both simulation of (possibly incomplete) speci cations and the generation of complete target applications (given code for the sequential computations of tasks). We have successfully used MASS in a case study involving several robots and a complex conveyer system.

Experience with a Portable Parallelizing Pascal Compiler

by Eran Gabber and Amiram Yehudai

International Conference on Parallel Processing, 1991

Page 223. 1991 International Conference on Parallel Processing Experience with a Portable Paralle... more

OBSERV/CRUISE - A CASE environment for prototyping of complex systems

by Amiram Yehudai and Shmuel S Tyszberowicz

. The paper presents a CASE environment for rapid prototyping of software. The environment suppor... more . The paper presents a CASE environment for rapid prototyping of software. The environment supports the OBSERVmethodologyand language,which combines the objectoriented approach (for system decomposition), finite state machines (to model behavior), and logic programming paradigm (for describing activities). The environment is based on a graphical front end, named CRUISE and includes tools for editing, simulation (or execution), browsing, and static checking. The paper outlines the OBSERV methodolgy, and presents in detail the CRUISE user interface, and each of the tools. 1 Introduction The traditional software development process is unsuitable for complex target systems (e.g. embedded systems). Rapid prototyping is increasingly seen as an important software development paradigm, augmenting or even replacing the conventional phased approach [3, 5, 13]. The OBSERV methodology for software development is based on rapid construction of an executable specification (or prototype) of ...

Locating Regression Bugs

by Shmuel S Tyszberowicz and Amiram Yehudai

Lecture Notes in Computer Science, 2008

ABSTRACT A regression bug is a bug which causes a feature that worked correctly to stop working a... more ABSTRACT A regression bug is a bug which causes a feature that worked correctly to stop working after a certain event (system upgrade, system patching, daylight saving time switch, etc.). Very often an encompassed bug fix included in a patch causes the regression bug. Regression bugs are an annoying and painful phenomena in the software development process, requiring a great deal of effort to find. Many tools have been developed in order to find the existence of such bugs. However, a great deal of manual work is still needed to find the exact source-code location that caused a regression bug. In this paper we present the CodePsychologist, a tool which assists the programmer to locate source code segments that caused a given regression bug. The CodePsychologist goes beyond current tools, that identify all the lines of code that changed since the feature in question worked properly (with the help of a Source Control Tool). The CodePsychologist uses various heuristics to select the lines most likely to be the cause of the error, from these often large number of lines of code. This reduces the fixing time of regression bugs. It allows a quick response to field errors that need an immediate correction.

KeYGenU: combining verification-based and capture and replay techniques for regression unit testing

by Amiram Yehudai and Shmuel S Tyszberowicz

International Journal of System Assurance Engineering and Management, 2011

Unit testing plays a major role in the software development process. Two essential criteria to ac... more Unit testing plays a major role in the software development process. Two essential criteria to achieve effective unit testing are: (1) testing each unit in isolation from other parts of the program and (2) achieving high code coverage. The former requires a lot of extra work such as writing drivers and stubs, whereas the latter is difficult to achieve when manually writing the tests. When changing existing code it is advocated to run the unit tests to avoid regression bugs. However, in many cases legacy software has no unit tests. Writing those tests from scratch is a hard and tedious process, which might not be cost-effective. This paper presents a tool chain approach that combines verification-based testing (VBT) and capture and replay (CaR) test generation methods. We have built a concrete tool chain, KeYGenU, which consists of two existing tools -KeY and GenUTest. The KeY sys-tem is a deductive verification and test-generation tool. GenUTest automatically generates JUnit tests for a correctly working software. This combination provides isolated unit test suites with high code-coverage. The generated tests can also be used for regression testing.

Generating Regression Unit Tests Using a Combination of Verification and Capture & Replay

by Shmuel S Tyszberowicz and Amiram Yehudai

Lecture Notes in Computer Science, 2010

The combination of software verification and testing techniques is increasingly encouraged due to... more The combination of software verification and testing techniques is increasingly encouraged due to their complementary strengths. Some verification tools have extensions for test case generation. These tests are strong at detecting software faults during the implementation and verification phase, and to further increase the confidence in the final software product. However, tests generated using verification technology alone may lack some of the benefits obtained by using more traditional testing techniques. One such technique is Capture and Replay, whose strengths are the generation of isolated unit tests and regression test oracles. Hence, the two groups of techniques have complementary strengths, and therefore are ideal candidates for a tool-chain approach proposed in this paper. The first phase produces, for a given system, unit tests with high coverage. However, when using them to test a unit, its environment is tested as well -resulting in a high cost of testing. To solve this problem, the second phase captures the various executions of the program, which are monitored by the output of the first phase. The output of the second phase is a set of unit tests with high code coverage, which uses mock objects to test the units. Another advantage of this approach is the fact that the generated tests can also be used for regression testing.

KeYGenU: combining verification-based and capture and replay techniques for regression unit testing

by Shmuel Tyszberowicz, Bernhard Beckert, and Amiram Yehudai

International Journal of System Assurance Engineering and Management, 2011

Unit testing plays a major role in the software development process. Two essential criteria to ac... more Unit testing plays a major role in the software development process. Two essential criteria to achieve effective unit testing are: (1) testing each unit in isolation from other parts of the program and (2) achieving high code coverage. The former requires a lot of extra work such as writing drivers and stubs, whereas the latter is difficult to achieve when manually writing the tests. When changing existing code it is advocated to run the unit tests to avoid regression bugs. However, in many cases legacy software has no unit tests. Writing those tests from scratch is a hard and tedious process, which might not be cost-effective. This paper presents a tool chain approach that combines verification-based testing (VBT) and capture and replay (CaR) test generation methods. We have built a concrete tool chain, KeYGenU, which consists of two existing tools -KeY and GenUTest. The KeY sys-tem is a deductive verification and test-generation tool. GenUTest automatically generates JUnit tests for a correctly working software. This combination provides isolated unit test suites with high code-coverage. The generated tests can also be used for regression testing.

Combining reflection and finite state diagrams for design enforcement

by Vered Gafni and Amiram Yehudai

Theory and Practice of Object Systems, 1996

This paper presents a method for design enforcement, based on a combination of re ection and Stat... more This paper presents a method for design enforcement, based on a combination of re ection and State Machine Diagrams. The method is supported by RECOS, a RE ective system in a Concurrent Object-oriented environment that uses Finite State Diagrams. The re ection methodology is now a developing branch in computer science for controlling and modifying systems, particularly Object-Oriented systems. RECOS is a \base-system" under which other systems can be built. Combining concepts of Concurrent Object-Oriented design, Finite State Diagrams, and re ection leads to an increase in the reliability of the systems, by insuring that objects function according to their design. RECOS also presents a safe way for handling modi cations of the systems. We attempt to provide the developer of a system under RECOS with an environment that balances between easing implementation e ort and guarding against unsafe implementations. We address here the issue of re ection, present the structure of RECOS, using as a running example a heating system, and brie y discuss implementation details and future work.

Reasoning About Real-Time Reactive Systems with MASS

by Yishai Feldman, Amiram Yehudai, and Vered Gafni

This paper lays the theoretical foundation for formal reasoning with the language MASS. MASS is a... more This paper lays the theoretical foundation for formal reasoning with the language MASS. MASS is an activation oriented speci cation language, suitable for specifying reactive real time systems. Following a short review of MASS, the paper presents a logical formalism based on the language of MASS events, for stating properties of MASS speci cations. A version of MASS in which speci ed reactions must be given explicit deadlines is then shown to be decidable. The general unbounded version is shown to be undecidable.

Formal Development of Real-time Systems with PLOT/MASS

by Yishai Feldman, Amiram Yehudai, and Vered Gafni

The paper presents a formal development process of a real-time system. We employ two related lang... more The paper presents a formal development process of a real-time system. We employ two related languages: the real-time logic plot for plant modeling and requirements speci cation, and the real-time language mass for the controller design. A deductive proof system is used to verify the design against the requirements. The examples given in the paper demonstrate the special features of the mass/plot framework that make it especially suitable for real-time systems speci cation. In particular, we show explicit expression of causality and timing in plot, modular design by re nement in mass, and the succinct and intuitive nature of speci cations in this framework.

Real-Time Systems Development with MASS??

by Yishai Feldman, Amiram Yehudai, and Vered Gafni

Electronic Notes in Theoretical Computer Science, 1999

In this paper, we demonstrate the capability of mass, a real-time design language, for large syst... more In this paper, we demonstrate the capability of mass, a real-time design language, for large systems speci cation. The paper presents a hierarchical speci cation of an automatic cruise controller that evolves through stepwise re nement. In particular, we show modular design, the separation of the functional and reactive concerns, and the succinct and intuitive nature of speci cations in mass.

Activation-oriented specification of real-time systems

by Yishai Feldman, Amiram Yehudai, and Vered Gafni

Lecture Notes in Computer Science, 1994

We p r e s e n t MASS, a declarative language for specifying the reactive b e h a vior of real-ti... more We p r e s e n t MASS, a declarative language for specifying the reactive b e h a vior of real-time systems. The basic primitive of the language is the task, which i s t h e i n terface between the reactive and sequential aspects of the speci cation. The purely computational meaning of a task (as an I/O transformation) is given outside MASS using standard speci cation languages for sequential computations. The reactive aspects of real-time systems are expressed in MASS through causal and temporal relations between events that signal task terminations. Hierarchical systems are obtained by re ning tasks, specifying them as sub-systems in MASS. MASS is given a formal semantics in a trace model augmented with explicit representation of causes. A synchronous execution model conforming with the formal semantics makes MASS speci cations executable. A development system based on this model allows both simulation of (possibly incomplete) speci cations and the generation of complete target applications (given code for the sequential computations of tasks). We have successfully used MASS in a case study involving several robots and a complex conveyer system.

Experience with a Portable Parallelizing Pascal Compiler

by Eran Gabber and Amiram Yehudai

International Conference on Parallel Processing, 1991

Page 223. 1991 International Conference on Parallel Processing Experience with a Portable Paralle... more