Skip to main content

Walter Tichy

Karlsruhe Institute of Technology (KIT), Faculty of Informatics, Faculty Member

Followers

36

Following

13

Co-authors

11

Public Views

Kolej Komuniti Bandar Darulaman

University of Oxford

Utkir Khamdamov

Tashkent University of Information Technologies

JAY- MELT Official

University of Edinburgh

ogirdor carvallaso

Interests

Uploads

Papers by Walter Tichy

Dagstuhl-Manifest zur strategischen Bedeutung des Software Engineering in Deutschland

by Volker Gruhn, M. Nagl, Lutz Prechelt, Andreas Rausch, Bernhard Rumpe, and Walter Tichy

Software ist der fundamentale Werkstoff des Informationszeitalters. Innovative Produkte und Diens... more Software ist der fundamentale Werkstoff des Informationszeitalters. Innovative Produkte und Dienstleistungen sind ohne Software nicht mehr denkbar. Die Wettbewerbsfähigkeit der deutschen Wirtschaft hängt entscheidend von der Fähigkeit ab, Software-intensive Produkte und Dienstleistungen mit höchster Qualität zu erstellen. Software Engineering über dem Weltniveau ist die Voraussetzung dafür, dass Deutschland seine führende Stellung im Ingenieurbereich, etwa im Maschinenbau, halten und ausbauen und entsprechende Positionen in neuen Sparten, etwa im modernen Gesundheitswesen (e-Health), aufbauen kann. Software wird in der Zukunft integrierter -in vielen Fällen sogar dominierender -Teil großer komplexer Systeme sein. Nicht nur in der Automobil-und Luftfahrtindustrie wird dieser Trend bereits heute deutlich sichtbar. Die erforderliche Integration von Mechanik, Elektronik und Software und die Vermeidung unerwünschter Wechselwirkungen kann nur durch die frühzeitige Integration der Modellierungskompetenzen des Software Engineering in den Entwicklungsprozess beherrscht werden. Diese neue Positionierung von Software Engineering als systemische Disziplin erfordert eine neue Ausrichtung und Stärkung der Bereiche Forschung, Lehre und Technologietransfer. Es ist die auf sorgfältige Analysen abgestützte Überzeugung der Unterzeichner, dass verstärkte Anstrengungen in allen drei Bereichen notwendig sind, um die Herausforderungen des Informationszeitalters und der Globalisierung anzunehmen. Dieses Manifest stellt Forderungen inhaltlicher und organisatorischer Art für Forschung, Lehre und Technologietransfer auf, deren Erfüllung Voraussetzung für die Realisierung künftiger weltweiter wirtschaftlicher Potentiale ist. Der Standort Deutschland wird sich in diesem Markt nur behaupten, wenn Software Engineering sowie Softwareingenieure auf höchstem Qualitäts-und Produktivitätsniveau verfügbar sind.

Dagstuhl-Manifest zur strategischen Bedeutung des Software Engineering in Deutschland

by Armin Cremers, Bernhard Rumpe, and Walter Tichy

Software hat sich zu einem zentralen Werkstoff des Informationszeitalters entwickelt. Innovative ... more Software hat sich zu einem zentralen Werkstoff des Informationszeitalters entwickelt. Innovative Produkte und Dienstleistungen sind ohne zuverlässige Software nicht mehr denkbar. Dementsprechend hängt die Wettbewerbsfähigkeit der deutschen Wirtschaft entscheidend von der Fähigkeit ab, Software-intensive Produkte und Dienstleistungen mit höchster Qualität schnell und kostengünstig zu erstellen. Software Engineering (abgekürzt SE, deutsch Softwaretechnik) auf Weltniveau ist die Voraussetzung dafür, dass Deutschland seine führende Stellung im Ingenieurbereich (z.B. Export-Weltmeister im Maschinenbau) halten und ausbauen bzw. eine entsprechende Position in neuen Sparten (z.B. e-Health) aufbauen kann. Die Bedeutung von Software Engineering hat sich verstärkt und wird weiter wachsen. Während Software in der Vergangenheit eine weitgehend isolierte Komponente großer komplexer Systeme war, so wird sie in der Zukunft integrierter -in vielen Fällen sogar dominierender -Teil großer komplexer Systeme sein. In der Automobilindustrie wird dieser Trend bereits heute deutlich sichtbar. Die erforderliche Integration von Mechanik, Elektronik und Software und die Vermeidung unerwünschter Wechselwirkungen kann nur durch die frühzeitige Integration der Modellierungskompetenzen des Software Engineering in den Entwicklungsprozess beherrscht werden. Diese neue Positionierung von Software Engineering als systemische Disziplin erfordert eine neue Ausrichtung in den drei Bereichen Forschung, Lehre und Technologietransfer. Dieses Manifest stellt Forderungen inhaltlicher und organisatorischer Art für Forschung, Lehre und Technologietransfer auf, deren Erfüllung der wirtschaftlichen Bedeutung von Software und Software Engineering Rechnung tragen soll. Es ist die auf sorgfältige Untersuchungen abgestützte Überzeugung der Unterzeichner, dass verstärkte Anstrengungen in allen drei Bereichen notwendig sind, um den Herausforderungen [BJN+06] M. Broy, M.

The TILDE Project

by John Korb, Walter Tichy, and Tim Korb

The TILDE project investigates general-purpose computing systems that run on a multi·machine comp... more The TILDE project investigates general-purpose computing systems that run on a multi·machine computing engine. A multi-machine computing engine is a cluster of heterogeneous processors loosely coupled with a high-speed local area network. The goal of the project is to explore computing systems in which the user interlace hides details of the underlying architecture, making the distributed computing engine appear to be a single, large time-sharing system. This paper presents prqject plans and status. It describes the architecture of the TILDE computing engine as well as research in the areas of operating systems. user interfaces, and computational services. The novel computational services planned for TILDE include high-level electronic mail, a timed event service, satellite bulletin· board broadcast, supercomputer access, and otbers.

Compiling for Massively Parallel Machines

Dagstuhl Seminars, 1991

This article discusses techniques for compiling high-level, explicitly-parallellanguages for mass... more This article discusses techniques for compiling high-level, explicitly-parallellanguages for massively parallel machines.We present mechanisms for translating asynchronous as well as synchronousparallelism for both SIMD and MIMD machines. We showhow the parallelism specified in a program is mapped onto the availableprocessors and discuss an effective optimization that eliminates redundantsynchronization points. Approaches for improving scheduling, loadbalancing, and co-location of ...

Summary of the Dagstuhl workshop on future directions in software engineering: February 17--21, 1992, Schloß Dagstuhl

ACM Sigsoft Software Engineering Notes, 1993

The intent of the workshop was to bring together leading scientists for identifying promising dir... more The intent of the workshop was to bring together leading scientists for identifying promising directions for future research in Software Engineering. The motivation for the workshop was the realization that Software Engineering research was not in good shape, with the present emphasis on management and risk control diverting attention from hard, technical subjects.After week-long, intensive discussions of a great number

An experiment to assess the benefits of inter-module type checking

Proceedings of the 3rd International Software Metrics Symposium, 1996

This paper reports on an experiment to assess the error detection capabilities of static, inter-m... more

The Experimental Paradigm in Software Engineering

Dagstuhl Seminars, 1992

This paper appears in Experimental Software Engineering Issues: Critical Assessment and Future Di... more This paper appears in Experimental Software Engineering Issues: Critical Assessment and Future Directives, Proceedings of Dagstuhl-Workshop, edited by H. Dieter Rombach, Victor R. Basili , and Richard Selby,, September 1992, published by Springer-Verlag, #706, Lecture Notes in Computer Software, August 1993. modeling research that does exist has centered on the software product, specifically mathematical models of the program function. We

Two Controlled Experiments Assessing the Usefulness of Design Pattern Information in Program Maintenance

IEEE Transactions on Software Engineering, 2000

A major hypothesis underlying design pattern research is th at use of design pat- terns improves ... more A major hypothesis underlying design pattern research is th at use of design pat- terns improves programmer productivity and software quality. This paper reports on a first experimental test of this hypothesis. Subjects performed maintenance tasks on two programs ranging from 360 to 560 LOC including comments. Both programs contained design patterns. The controlled variable was whether the design patterns

A Controlled Experiment in Maintenance Comparing Design Patterns to Simpler Solutions

IEEE Transactions on Software Engineering, 1998

Software design patterns package proven solutionsto recurring design problems in a formthat simpl... more Software design patterns package proven solutionsto recurring design problems in a formthat simplifies reuse. We are seeking empiricalevidence whether using design patterns is beneficial.In particular, one may prefer using adesign pattern even if the actual design problemis simpler than that solved by the pattern,i.e., if not all of the functionality offered by thepattern is actually required.Our experiment investigates software maintenancescenarios

Summary of the Dagstuhl workshop on software architecture

by Walter Tichy and Frances Paulisch

ACM SIGSOFT Software Engineering Notes, 1995

... Siemens - Princeton, NJ, USA Walter Tichy ... Legacy Architecture Dewayne E. Perry AT&amp... more

Atune-IL: An Instrumentation Language for Auto-tuning Parallel Applications

Lecture Notes in Computer Science, 2009

Auto-tuners automate the performance tuning of parallel applications. Three major drawbacks of cu... more Auto-tuners automate the performance tuning of parallel applications. Three major drawbacks of current approaches are 1) they mainly focus on numerical software; 2) they typically do not attempt to reduce the large search space before search algorithms are applied; 3) the means to provide an auto-tuner with additional information to improve tuning are limited. Our paper tackles these problems in

Renaming Detection

by Walter Tichy and James Hunt

Finding changed identifiers in programs is important for program comparison and merging. Comparin... more Finding changed identifiers in programs is important for program comparison and merging. Comparing two versions of a program is complicated if renaming has occurred. Textual merging is highly unreliable if, in one version, identifiers were renamed, while in the other version, code using the old identifiers was added or modified.

Delta Algorithms: An empirical evaluation

ACM Transactions on Software Engineering and Methodology, 1998

Truck Scheduling on Multicore

it - Information Technology, 2011

Extensible language-aware merging

International Conference on Software Maintenance, 2002. Proceedings., 2002

Parallel development has become standard practice in software development and maintenance. Though... more Parallel development has become standard practice in software development and maintenance. Though most every revision control and configuration management system provides some form of merging for combining changes made in parallel, these mechanisms often yield unsatisfactory results. The authors present a new merging algorithm, that uses a fast differencing algorithm and renaming analysis to provide better merge results. The system is language aware, but not language dependent and does not require a special editor, so it can be easily integrated in current development environments.

Delta algorithms: an empirical analysis

ACM Transactions on Software Engineering and Methodology, 1998

Delta algorithms compress data by encoding one file in terms of another. This type of compression... more Delta algorithms compress data by encoding one file in terms of another. This type of compression is useful in a number of situations: strong multiple versions of data, displaying differences, merging changes, distributing updates, storing backups, transmitting video sequences, and others. This article studies the performance parameters of several delta algorithms, using a benchmark of over 1,300 pairs of files

An Empirical Study of Delta Algorithms

System Configuration Management, 1996

Delta algorithms compress data by encoding one le in terms of another. This type of compression i... more Delta algorithms compress data by encoding one le in terms of another. This type of compression is useful in a number of situations: storing multiple versions of data, distributing updates, storing backups, transmitting video sequences, and others. This paper studies the performance parameters of several delta algorithms, using a benchmark of over 1300 pairs of les taken from two successive releases of GNU software. Results indicate that modern delta compression algorithms based on Ziv-Lempel techniques signi cantly outperform di , a popular but older delta compressor, in terms of compression ratio. The modern compressors also correlate better with the actual di erence between les; one of them is even faster than di in both compression and decompression speed.

An Empirical Study of Delta Mechanisms

Delta algorithms compress data by encoding one file in terms of another. This type of compression... more Delta algorithms compress data by encoding one file in terms of another. This type of compression is useful in a number of situations: storing multiple versions of data, displaying differences, merging changes, distributing updates, storing backups, transmitting video sequences, and others. This article studies the performance parameters of several delta algorithms, using a benchmark of over 1,300 pairs of files taken from two successive releases of GNU software. Results indicate that modern delta compression algorithms based on Ziv-Lempel techniques significantly outperform diff, a popular but older delta compressor, in terms of compression ratio. The modern compressors also correlate better with the actual difference between files without sacrificing performance.

Edge: An extendible graph editor

by Walter Tichy and Frances Paulisch

Software: Practice and Experience, 1990

EDGE is an editor kernel for the direct and visual manipulation of graphs. The kernel can be adap... more EDGE is an editor kernel for the direct and visual manipulation of graphs. The kernel can be adapted quickly to diverse applications based on graphs, such as PERT chart editing, directory browsing, call graph display, logic circuit simulation or configuration visualization. EDGE provides potential solutions to the following general problems faced by any graph editor. (1) Automatic graph layout: how can application-specific layout requirements, individual preferences, and layout stability be integrated with automatic layout algorithms? EDGE solves this problem with a novel algorithm that is based on layout constraints. (2) Graph abstraction: how can users deal with large graphs containing hundreds of nodes and edges, and thousands of edge crossings? EDGE can reduce the apparent complexity with subgraph abstractions and a novel clustering technique called edge concentration. (3) Adaptability: how should the editor kernel be structured to be adaptable to various applications? EDGE uses a special graph representation language for specifying visual appearance and the inheritance mechanism of C+ + to achieve extendibility. (4) Persistence of graphs: how can the graph structures produced by the editor be kept in long-term storage, especially if the node and edge data structures have been extended for a particular application? Our approach uses a standardized, external format for graphs and handles extensions with program generator technology: the I/O routines for reading and writing extended node and edge data structures are produced automatically from the declarations of these data structures. This paper describes EDGE and presents details of the above solutions.

Experimental evaluation in computer science: A quantitative study

by Lutz Prechelt, Walter Tichy, and Ernst Heinz

Journal of Systems and Software, 1995

A survey of over 400 recent research articles suggests that computer scientists publish relativel... more

Dagstuhl-Manifest zur strategischen Bedeutung des Software Engineering in Deutschland

by Volker Gruhn, M. Nagl, Lutz Prechelt, Andreas Rausch, Bernhard Rumpe, and Walter Tichy

Software ist der fundamentale Werkstoff des Informationszeitalters. Innovative Produkte und Diens... more Software ist der fundamentale Werkstoff des Informationszeitalters. Innovative Produkte und Dienstleistungen sind ohne Software nicht mehr denkbar. Die Wettbewerbsfähigkeit der deutschen Wirtschaft hängt entscheidend von der Fähigkeit ab, Software-intensive Produkte und Dienstleistungen mit höchster Qualität zu erstellen. Software Engineering über dem Weltniveau ist die Voraussetzung dafür, dass Deutschland seine führende Stellung im Ingenieurbereich, etwa im Maschinenbau, halten und ausbauen und entsprechende Positionen in neuen Sparten, etwa im modernen Gesundheitswesen (e-Health), aufbauen kann. Software wird in der Zukunft integrierter -in vielen Fällen sogar dominierender -Teil großer komplexer Systeme sein. Nicht nur in der Automobil-und Luftfahrtindustrie wird dieser Trend bereits heute deutlich sichtbar. Die erforderliche Integration von Mechanik, Elektronik und Software und die Vermeidung unerwünschter Wechselwirkungen kann nur durch die frühzeitige Integration der Modellierungskompetenzen des Software Engineering in den Entwicklungsprozess beherrscht werden. Diese neue Positionierung von Software Engineering als systemische Disziplin erfordert eine neue Ausrichtung und Stärkung der Bereiche Forschung, Lehre und Technologietransfer. Es ist die auf sorgfältige Analysen abgestützte Überzeugung der Unterzeichner, dass verstärkte Anstrengungen in allen drei Bereichen notwendig sind, um die Herausforderungen des Informationszeitalters und der Globalisierung anzunehmen. Dieses Manifest stellt Forderungen inhaltlicher und organisatorischer Art für Forschung, Lehre und Technologietransfer auf, deren Erfüllung Voraussetzung für die Realisierung künftiger weltweiter wirtschaftlicher Potentiale ist. Der Standort Deutschland wird sich in diesem Markt nur behaupten, wenn Software Engineering sowie Softwareingenieure auf höchstem Qualitäts-und Produktivitätsniveau verfügbar sind.

Dagstuhl-Manifest zur strategischen Bedeutung des Software Engineering in Deutschland

by Armin Cremers, Bernhard Rumpe, and Walter Tichy

Software hat sich zu einem zentralen Werkstoff des Informationszeitalters entwickelt. Innovative ... more Software hat sich zu einem zentralen Werkstoff des Informationszeitalters entwickelt. Innovative Produkte und Dienstleistungen sind ohne zuverlässige Software nicht mehr denkbar. Dementsprechend hängt die Wettbewerbsfähigkeit der deutschen Wirtschaft entscheidend von der Fähigkeit ab, Software-intensive Produkte und Dienstleistungen mit höchster Qualität schnell und kostengünstig zu erstellen. Software Engineering (abgekürzt SE, deutsch Softwaretechnik) auf Weltniveau ist die Voraussetzung dafür, dass Deutschland seine führende Stellung im Ingenieurbereich (z.B. Export-Weltmeister im Maschinenbau) halten und ausbauen bzw. eine entsprechende Position in neuen Sparten (z.B. e-Health) aufbauen kann. Die Bedeutung von Software Engineering hat sich verstärkt und wird weiter wachsen. Während Software in der Vergangenheit eine weitgehend isolierte Komponente großer komplexer Systeme war, so wird sie in der Zukunft integrierter -in vielen Fällen sogar dominierender -Teil großer komplexer Systeme sein. In der Automobilindustrie wird dieser Trend bereits heute deutlich sichtbar. Die erforderliche Integration von Mechanik, Elektronik und Software und die Vermeidung unerwünschter Wechselwirkungen kann nur durch die frühzeitige Integration der Modellierungskompetenzen des Software Engineering in den Entwicklungsprozess beherrscht werden. Diese neue Positionierung von Software Engineering als systemische Disziplin erfordert eine neue Ausrichtung in den drei Bereichen Forschung, Lehre und Technologietransfer. Dieses Manifest stellt Forderungen inhaltlicher und organisatorischer Art für Forschung, Lehre und Technologietransfer auf, deren Erfüllung der wirtschaftlichen Bedeutung von Software und Software Engineering Rechnung tragen soll. Es ist die auf sorgfältige Untersuchungen abgestützte Überzeugung der Unterzeichner, dass verstärkte Anstrengungen in allen drei Bereichen notwendig sind, um den Herausforderungen [BJN+06] M. Broy, M.

The TILDE Project

by John Korb, Walter Tichy, and Tim Korb

The TILDE project investigates general-purpose computing systems that run on a multi·machine comp... more The TILDE project investigates general-purpose computing systems that run on a multi·machine computing engine. A multi-machine computing engine is a cluster of heterogeneous processors loosely coupled with a high-speed local area network. The goal of the project is to explore computing systems in which the user interlace hides details of the underlying architecture, making the distributed computing engine appear to be a single, large time-sharing system. This paper presents prqject plans and status. It describes the architecture of the TILDE computing engine as well as research in the areas of operating systems. user interfaces, and computational services. The novel computational services planned for TILDE include high-level electronic mail, a timed event service, satellite bulletin· board broadcast, supercomputer access, and otbers.

Compiling for Massively Parallel Machines

Dagstuhl Seminars, 1991

This article discusses techniques for compiling high-level, explicitly-parallellanguages for mass... more This article discusses techniques for compiling high-level, explicitly-parallellanguages for massively parallel machines.We present mechanisms for translating asynchronous as well as synchronousparallelism for both SIMD and MIMD machines. We showhow the parallelism specified in a program is mapped onto the availableprocessors and discuss an effective optimization that eliminates redundantsynchronization points. Approaches for improving scheduling, loadbalancing, and co-location of ...

Summary of the Dagstuhl workshop on future directions in software engineering: February 17--21, 1992, Schloß Dagstuhl

ACM Sigsoft Software Engineering Notes, 1993

The intent of the workshop was to bring together leading scientists for identifying promising dir... more The intent of the workshop was to bring together leading scientists for identifying promising directions for future research in Software Engineering. The motivation for the workshop was the realization that Software Engineering research was not in good shape, with the present emphasis on management and risk control diverting attention from hard, technical subjects.After week-long, intensive discussions of a great number

An experiment to assess the benefits of inter-module type checking

Proceedings of the 3rd International Software Metrics Symposium, 1996

This paper reports on an experiment to assess the error detection capabilities of static, inter-m... more

The Experimental Paradigm in Software Engineering

Dagstuhl Seminars, 1992

This paper appears in Experimental Software Engineering Issues: Critical Assessment and Future Di... more This paper appears in Experimental Software Engineering Issues: Critical Assessment and Future Directives, Proceedings of Dagstuhl-Workshop, edited by H. Dieter Rombach, Victor R. Basili , and Richard Selby,, September 1992, published by Springer-Verlag, #706, Lecture Notes in Computer Software, August 1993. modeling research that does exist has centered on the software product, specifically mathematical models of the program function. We

Two Controlled Experiments Assessing the Usefulness of Design Pattern Information in Program Maintenance

IEEE Transactions on Software Engineering, 2000

A major hypothesis underlying design pattern research is th at use of design pat- terns improves ... more A major hypothesis underlying design pattern research is th at use of design pat- terns improves programmer productivity and software quality. This paper reports on a first experimental test of this hypothesis. Subjects performed maintenance tasks on two programs ranging from 360 to 560 LOC including comments. Both programs contained design patterns. The controlled variable was whether the design patterns

A Controlled Experiment in Maintenance Comparing Design Patterns to Simpler Solutions

IEEE Transactions on Software Engineering, 1998

Software design patterns package proven solutionsto recurring design problems in a formthat simpl... more Software design patterns package proven solutionsto recurring design problems in a formthat simplifies reuse. We are seeking empiricalevidence whether using design patterns is beneficial.In particular, one may prefer using adesign pattern even if the actual design problemis simpler than that solved by the pattern,i.e., if not all of the functionality offered by thepattern is actually required.Our experiment investigates software maintenancescenarios

Summary of the Dagstuhl workshop on software architecture

by Walter Tichy and Frances Paulisch

ACM SIGSOFT Software Engineering Notes, 1995

... Siemens - Princeton, NJ, USA Walter Tichy ... Legacy Architecture Dewayne E. Perry AT&amp... more

Atune-IL: An Instrumentation Language for Auto-tuning Parallel Applications

Lecture Notes in Computer Science, 2009

Auto-tuners automate the performance tuning of parallel applications. Three major drawbacks of cu... more Auto-tuners automate the performance tuning of parallel applications. Three major drawbacks of current approaches are 1) they mainly focus on numerical software; 2) they typically do not attempt to reduce the large search space before search algorithms are applied; 3) the means to provide an auto-tuner with additional information to improve tuning are limited. Our paper tackles these problems in

Renaming Detection

by Walter Tichy and James Hunt

Finding changed identifiers in programs is important for program comparison and merging. Comparin... more Finding changed identifiers in programs is important for program comparison and merging. Comparing two versions of a program is complicated if renaming has occurred. Textual merging is highly unreliable if, in one version, identifiers were renamed, while in the other version, code using the old identifiers was added or modified.

Delta Algorithms: An empirical evaluation

ACM Transactions on Software Engineering and Methodology, 1998

Truck Scheduling on Multicore

it - Information Technology, 2011

Extensible language-aware merging

International Conference on Software Maintenance, 2002. Proceedings., 2002

Parallel development has become standard practice in software development and maintenance. Though... more Parallel development has become standard practice in software development and maintenance. Though most every revision control and configuration management system provides some form of merging for combining changes made in parallel, these mechanisms often yield unsatisfactory results. The authors present a new merging algorithm, that uses a fast differencing algorithm and renaming analysis to provide better merge results. The system is language aware, but not language dependent and does not require a special editor, so it can be easily integrated in current development environments.

Delta algorithms: an empirical analysis

ACM Transactions on Software Engineering and Methodology, 1998

Delta algorithms compress data by encoding one file in terms of another. This type of compression... more Delta algorithms compress data by encoding one file in terms of another. This type of compression is useful in a number of situations: strong multiple versions of data, displaying differences, merging changes, distributing updates, storing backups, transmitting video sequences, and others. This article studies the performance parameters of several delta algorithms, using a benchmark of over 1,300 pairs of files

An Empirical Study of Delta Algorithms

System Configuration Management, 1996

Delta algorithms compress data by encoding one le in terms of another. This type of compression i... more Delta algorithms compress data by encoding one le in terms of another. This type of compression is useful in a number of situations: storing multiple versions of data, distributing updates, storing backups, transmitting video sequences, and others. This paper studies the performance parameters of several delta algorithms, using a benchmark of over 1300 pairs of les taken from two successive releases of GNU software. Results indicate that modern delta compression algorithms based on Ziv-Lempel techniques signi cantly outperform di , a popular but older delta compressor, in terms of compression ratio. The modern compressors also correlate better with the actual di erence between les; one of them is even faster than di in both compression and decompression speed.

An Empirical Study of Delta Mechanisms

Delta algorithms compress data by encoding one file in terms of another. This type of compression... more Delta algorithms compress data by encoding one file in terms of another. This type of compression is useful in a number of situations: storing multiple versions of data, displaying differences, merging changes, distributing updates, storing backups, transmitting video sequences, and others. This article studies the performance parameters of several delta algorithms, using a benchmark of over 1,300 pairs of files taken from two successive releases of GNU software. Results indicate that modern delta compression algorithms based on Ziv-Lempel techniques significantly outperform diff, a popular but older delta compressor, in terms of compression ratio. The modern compressors also correlate better with the actual difference between files without sacrificing performance.

Edge: An extendible graph editor

by Walter Tichy and Frances Paulisch

Software: Practice and Experience, 1990

EDGE is an editor kernel for the direct and visual manipulation of graphs. The kernel can be adap... more EDGE is an editor kernel for the direct and visual manipulation of graphs. The kernel can be adapted quickly to diverse applications based on graphs, such as PERT chart editing, directory browsing, call graph display, logic circuit simulation or configuration visualization. EDGE provides potential solutions to the following general problems faced by any graph editor. (1) Automatic graph layout: how can application-specific layout requirements, individual preferences, and layout stability be integrated with automatic layout algorithms? EDGE solves this problem with a novel algorithm that is based on layout constraints. (2) Graph abstraction: how can users deal with large graphs containing hundreds of nodes and edges, and thousands of edge crossings? EDGE can reduce the apparent complexity with subgraph abstractions and a novel clustering technique called edge concentration. (3) Adaptability: how should the editor kernel be structured to be adaptable to various applications? EDGE uses a special graph representation language for specifying visual appearance and the inheritance mechanism of C+ + to achieve extendibility. (4) Persistence of graphs: how can the graph structures produced by the editor be kept in long-term storage, especially if the node and edge data structures have been extended for a particular application? Our approach uses a standardized, external format for graphs and handles extensions with program generator technology: the I/O routines for reading and writing extended node and edge data structures are produced automatically from the declarations of these data structures. This paper describes EDGE and presents details of the above solutions.

Experimental evaluation in computer science: A quantitative study

by Lutz Prechelt, Walter Tichy, and Ernst Heinz

Journal of Systems and Software, 1995

A survey of over 400 recent research articles suggests that computer scientists publish relativel... more