I'm a researcher and lecturer on Software Engineering related topics at Technische Universität Darmstadt, Germany. Furthermore, I'm a member of the expert-team on software architecture and software modernization of the gemeinnützige Gesellschaft zur Förderung des Forschungstransfers (GFFT e.V.).
My primary research focus is on static code analysis for software architecture analysis and code reviews. Additionally, I also do research related to software product lines, aspect-oriented programming, middleware and enterprise application design. [ACM author page]
My teaching focus is on software engineering related topics. At the bachelor level I regularly teach the course "Introduction to Software Engineering". At the Master's level I give lectures on "Component Technologies for Distributed Applications" and "Software Engineering Design & Construction". In the past, I also taught courses on "Databases I”, "Distributed Systems" and “Design Patterns” at the Duale Hochschule Baden-Württemberg, Mannheim.
I'm a "technology enthusiast" and regularly explore new technologies for the development of software applications. I'm particularly interested in exploring programming languages and software engineering tools.
Tom Dinkelaker, Michael Eichberg and Mira Mezini
Science of Computer Programming
Embedded domain-specific languages (EDSLs) are known to improve the productivity of developers. However, for many domains no DSL implementation is available and two important reasons for this are: First, the effort to implement EDSLs that provide the domain’s established syntax (called concrete syntax) is very high. Second, the EDSL and its underlying general-purpose programming language (GPL) are typically tightly integrated. This hampers reusability across different GPLs. Besides these implementation issues, the productivity gains of using EDSLs are also limited by the lack of explicit tool support for EDSL users—such as syntax highlighting or code analyses.
In this paper, we present an approach that significantly reduces the necessary effort to implement embedded DSLs with concrete syntax. The idea is to use island grammars to specify the EDSL’s concrete syntax. This enables the developer to implement the embedded DSL as a library and to incrementally specify the concrete syntax using meta-data. Only those parts of the EDSL’s grammar need to be specified that deviate from the grammar of the GPL. By analyzing an EDSL’s implementation using reflection, it is possible to provide tool support for EDSLs without having the developer implement it explicitly, such as syntax highlighting. An evaluation demonstrates the feasibility of our approach by embedding a real-world DSL into a GPL.
Martin Monperrus, Michael Eichberg, Elif Tekes and Mira Mezini
Empirical Software Engineering
Application Programming Interfaces (API) are exposed to developers in order to reuse software libraries. API directives are natural-language statements in API documentation that make developers aware of constraints and guidelines related to the usage of an API. This paper presents the design and the results of an empirical study on the directives of API documentation of object-oriented libraries. Its main contribution is to propose and extensively discuss a taxonomy of 23 kinds of API directives.
Thorsten Schäfer, Michael Eichberg, Michael Haupt and Mira Mezini
IEEE Transactions on Software Engineering
In this paper, we discuss a set of functional requirements for software exploration tools and provide initial evidence that various combinations of these features are needed to effectively assist developers in understanding software. We observe that current tools for software exploration only partly support these features. This has motivated the development of SEXTANT, a software exploration tool tightly integrated into the Eclipse IDE that has been developed to fill this gap. By means of case studies, we demonstrate how the requirements fulfilled by SEXTANT are conducive to an understanding needed to perform a maintenance task
Ralf Mitschke, Michael Eichberg, M. Mezini, Alessandro Garcia, Isela Macia
Proceedings of the 12th annual international conference on Aspect-oriented software development (AOSD 2013)
Checking a software's structural dependencies is a line of research on methods and tools for analyzing, modeling and checking the conformance of source code w.r.t. specifications of its intended static structure. Existing approaches have focused on the correctness of the specification, the impact of the approaches on software quality and the expressiveness of the modeling languages. However, large specifications become unmaintainable in the event of evolution without the means to modularize such specifications. We present Vespucci, a novel approach and tool that partitions a specification of the expected and allowed dependencies into a set of cohesive slices. This facilitates modular reasoning and helps individual maintenance of each slice. Our approach is suited for modeling high-level as well as detailed low-level decisions related to the static structure and combines both in a single modeling formalism. To evaluate our approach we conducted an extensive study spanning nine years of the evolution of the architecture of the object-relational mapping framework Hibernate.
Paolo G. Giarrusso, Klaus Ostermann, Michael Eichberg, Ralf Mitschke, Tillmann Rendel, Christian Kästner
Proceedings of the 12th annual international conference on Aspect-oriented software development (AOSD 2013)
Modularity and efficiency are often contradicting requirements, such that programers have to trade one for the other. We analyze this dilemma in the context of programs operating on collections. Performance-critical code using collections need often to be hand-optimized, leading to non-modular, brittle, and redundant code. In principle, this dilemma could be avoided by automatic collection-specific optimizations, such as fusion of collection traversals, usage of indexing, or reordering of filters. Unfortunately, it is not obvious how to encode such optimizations in terms of ordinary collection APIs, because the program operating on the collections is not reified and hence cannot be analyzed.
We propose SQuOpt, the Scala Query Optimizer--a deep embedding of the Scala collections API that allows such analyses and optimizations to be defined and executed within Scala, without relying on external tools or compiler extensions. SQuOpt provides the same "look and feel" (syntax and static typing guarantees) as the standard collections API. We evaluate SQuOpt by re-implementing several code analyses of the FindBugs tool using SQuOpt, show average speedups of 12x with a maximum of 12800x and hence demonstrate that SQuOpt can reconcile modularity and efficiency in real-world applications.
Technical Communications of the 27th Int’l. Conference on Logic Programming (ICLP’11)
Leibniz International Proceedings in Informatics (LIPIcs); 2011
Today, Prolog is often used to solve well-defined, domain-specific problems that are part of larger applications. In such cases, a tight integration of the Prolog program and the rest of the application, which is commonly written in a different language, is necessary. One common approach is to compile the Prolog code to (native) code in the target language. In this case, the effort necessary to build, test and deploy the final application is reduced.
However, most of the approaches that achieve reasonable performance compile Prolog to object-oriented code that relies on some kind of virtual machine (VM). These VMs are libraries implemented in the target language and implement Prolog’s execution semantics. This adds a significant layer to the object-oriented program and results in code that does not look and feel native to developers of object-oriented programs. Further, if Prolog’s execution semantics is implemented as a library the potential of modern runtime environments for object-oriented programs, such as the Java Virtual Machine, to effectively optimize the program is more limited. In this paper, we report on our approach to compile Prolog to high-level, idiomatic object-oriented Java code. The generated Java code closely resembles code written by Java developers and is effectively optimized by the Java Virtual Machine.
Tom Dinkelaker, Michael Eichberg and Mira Mezini
26th Symposium On Applied Computing
Embedded domain-specific languages (EDSLs) are known to improve the productivity of developers. However, for many domains no DSL implementation is available. Two important reasons are: First, the effort to implement embedded DSLs that provide the domain’s established syntax (called concrete syntax) is very high. Second, the embedded DSL and its underlying general-purpose programming language (GPL) are typically tightly integrated which hampers reusability across different GPLs. In this paper, we present an approach that significantly reduces the necessary effort to implement embedded DSLs with concrete syntax. The idea is to use island grammars to specify the EDSL’s concrete syntax. This enables the developer to implement the embedded DSL as a library and to incrementally specify the concrete syntax using meta-data. Only those parts of the EDSL’s grammar need to be specified that deviate from the grammar of the GPL and which is required to enable the integration with the GPL.
Michael Eichberg, Martin Monperrus, Sven Kloppenburg and Mira Mezini
Sixth European Conference on Modelling Foundations and Applications
Implementing static analyses of machine-level executable code is labor intensive and complex. We show how to leverage model-driven engineering to facilitate the design and implementation of programs doing static analyses. Further, we report on important lessons learned on the benefits and drawbacks while using the following technologies: using the Scala programming language as target of code generation, using XML-Schema to express a metamodel, using XSLT to implement transformations and for implementing a lint like tool. Finally, we report on the use of Prolog for writing model transformations.
Michael Eichberg, Karl Klose, Ralf Mitschke and Mira Mezini
13th International Symposium on Component Based Software Engineering
In general, components provide and require services and two components are bound if the first component provides a service required by the second component. However, certain variability in services – w.r.t. how and which functionality is provided or required – cannot be described using standard interface description languages. If this variability is relevant when selecting a matching component then human interaction is required to decide which components can be bound. We propose to use feature models for making this variability explicit and (re-)enabling automatic component binding. In our approach, feature models are one part of service specifications. This enables to declaratively specify which service variant is provided by a component. By referring to a service’s variation points, a component that requires a specific service can list the requirements on the desired variant. Using these specifications, a component environment can then determine if a binding of the components exists that satisfies all requirements. The prototypical environment Columbus demonstrates the feasibility of the approach.
Tom Dinkelaker, Michael Eichberg and Mira Mezini
Proceedings of 9th International Conference on Aspect-Oriented Software Development (AOSD)
ACM Press; 2010
Embedded domain-specific languages (EDSLs) are said to be easier to compose than DSLs that are implemented by preprocessors. However, existing approaches focus on composition scenarios where the use of abstractions from one do- main does not affect the interpretation of abstractions from another domain. This leads to programs that exhibit scattering and tangling symptoms if multiple EDSLs with cross-cutting domain semantics are used. To address this issue, we propose an architecture for embedding DSLs that makes use of meta-object protocols and aspect-oriented concepts to support crosscutting composition of EDSLs. This enables to write modularized EDSL programs where each program addresses one concern.
Michael Eichberg and Sven Kloppenburg and Karl Klose and Mira Mezini
Proceedings of International Conference on Software Engineering (ICSE)
ACM Press; 2008
Dependencies between program elements need to be modeled from diﬀerent perspectives reﬂecting architectural, design, and implementation level decisions. To avoid erosion of the intended structure of the code, it is necessary to explicitly codify these diﬀerent perspectives on the permitted dependencies and to detect violations continuously and incrementally as software evolves. We propose an approach that uses declarative queries to group source elements — across programming language module boundaries — into overlapping ensembles. The dependencies between these ensembles are also speciﬁed as logic queries. The approach has been integrated into the incremental build process of Eclipse to ensure continuous checking, using an engine for tabled and incremental evaluation of logic queries. Our evaluation shows that our approach is fast enough for day-to-day use along the incremental build process of modern IDEs.
Michael Eichberg and Matthias Kahl and Diptikalyan Saha and Mira Mezini and Klaus Ostermann
Proceedings of Practical Aspects of Declarative Languages, 9th International Symposium (PADL)
Springer [Lecture Notes in Computer Science (LNCS)](vol. 4354); 2007
Modern development environments integrate various static analyses into the build process. Analyses that analyze the whole project whenever the project changes are impractical in this context. We present an approach to automatic incrementalization of analyses that are specified as tabled logic programs and evaluated using incremental tabled evaluation, a technique for efficiently updating memo tables in response to changes in facts and rules. The approach has been implemented and integrated into the Eclipse IDE. Our measurements show that this technique is effective for automatically incrementalizing a broad range of static analyses.
Michael Eichberg and Mira Mezini and Sven Kloppenburg and Klaus Ostermann and Benjamin Rank
Proceedings of the 21st IEEE/ACM International Conference on Automated Software Engineering (ASE)
IEEE Computer Society; 2006
To improve the productivity of the development process, more and more tools for static software analysis are tightly integrated into the incremental build process of an IDE. If multiple interdependent analyses are used simultaneously, the coordination between the analyses becomes a major obstacle to keep the set of analyses open. We propose an approach to integrating and scheduling an open set of static analyses which decouples the individual analyses and coordinates the analysis executions such that the overall time and space consumption is minimized. The approach has been implemented for the Eclipse IDE and has been used to integrate a wide range of analyses such as finding bug patterns, detecting violations of design guidelines, or type system extensions for Java.
Michael Eichberg and Daniel Germanus and Mira Mezini and Lukas Mrokon and Thorsten Schäfer
Proceedings of the Conference on Software Maintenance and Reengineering (CSMR)
IEEE Computer Society; 2006
To measure the particularities of modern software development projects that use different types of documents for the implementation of a program, new metrics need to be defined. Further, well established metrics, such as e.g., lack of cohesion or coupling between objects need to be reconsidered in the presence of new language features. Not being able to thoroughly measure a project can lead to false conclusions with respect to the measured source files. Currently, a large number of metrics tools exists, but unfortunately most tools are not extensible, or they are limited with respect to the types of documents that can be taken into account. Further, support for testing newly developed metrics is also missing. In this paper we present QScope - an open, extensible metrics framework. QScope is open with respect to the supported artifacts and explicitly enables the user to implement new metrics by reasoning over all artifacts using a declarative query language. As we will show in this paper, using a declarative query language enables a concise definition of new metrics.
Michael Eichberg and Mira Mezini and Michael Haupt and Thorsten Schäfer
Proceedings of the 21st IEEE International Conference on Software Maintenance (ICSM)
IEEE Computer Society; 2005
Current tools for software understanding mostly concentrate on one comprehension technique, e.g., visualization, or bottom-up navigation through software elements via hyperlinks. In this paper, we argue that to effectively assist developers in understanding today's software systems, a combination of several comprehension techniques is needed, including seamless integration of top-down querying and bottom-up navigation strategies that work across different kinds of software artifacts; furthermore, application-domain and/or technology specific relationships between software elements should be taken into consideration; last but not least, a tight integration of such tools into development environments is crucial. We present Sextant, a software exploration tool tightly integrated into the Eclipse IDE that satisfies these requirements. In two case studies, we demonstrate how Sextant's features are conducive in tracking down the source of erroneous behavior, respectively, in discovering "bad smells" in the software structure which should lead to code refactorings.
Michael Eichberg and Thorsten Schäfer and Mira Mezini
Proceedings of Fundamental Approaches to Software Engineering: 8th International Conference (FASE)
Springer [Lecture Notes in Computer Science (LNCS)](vol. 3442); 2005
The specification of meta-information, by using attributes in .NET or annotations in Java, along with the source code is gaining widespread use. Meta-information is used for different purposes such as code generation or configuration of the environment in which a class is deployed. However, in most cases using an annotation also implies that constraints, beyond those defined by the language's semantics, have to be followed. E.g., a class must define a no-arguments constructor or the parameters of a method must have specific types. Currently, these constraints are not checked at all or only to a very limited extend. Hence, a violation can remain undetected and result in deployment-time or even subtle run-time errors. In this paper, we present a user-extensible framework that enables the definition of constraints to check the properties of annotated elements. Further, we demonstrate the application of the framework to check the constraints defined in the EJB 3.0 specification, and an evaluation of the approach based on checking the xPetstore-EJB3.0 project from within Eclipse to test the performance.
Michael Haupt, Mira Mezini, Christoph Bockisch, Tom Dinkelaker, Michael Eichberg, Michael Krebs
Proceedings of the 1st ACM/USENIX international conference on Virtual execution environments
Language mechanisms deserve language implementation effort. While this maxim has led to sophisticated support for language features specific to object-oriented, functional and logic programming languages, aspect-oriented programming languages are still mostly implemented using postprocessors. The Steamloom virtual machine, based on IBM's Jikes RVM, provides support for aspect-oriented programming at virtual machine level. A bytecode framework called BAT was integrated with the Jikes RVM to replace its bytecode management logic. While preserving the functionality needed by the VM, BAT also allows for querying application code for join point shadows, avoiding redundancy in bytecode representation. Performance measurements show that an AOP-enabled virtual machine like Steamloom does not inflict unnecessary performance penalties on a running application; when it comes to executing AOP-related operations, there even are significant performance gains compared to other approaches.
Michael Eichberg and Mira Mezini and Klaus Ostermann and Thorsten Schäfer
Proceedings of the 11th Working Conference on Reverse Engineering (WCRE)
IEEE Computer Society; 2004
We describe XIRC, a tool and architecture that enables to define queries over a uniform representation of all artifacts of a software project. These queries can be used for general cross-artifact information retrieval or for more special applications like checking implementation restrictions or conformance to style guides. XIRC is also a good basis to implement a broad range of tools for refactoring, generators, aspect-oriented programming and many other domains on top of it.
Michael Eichberg and Mira Mezini and Klaus Ostermann
Proceedings of Programming Languages and Systems: Second Asian Symposium (APLAS)
Springer [Lecture Notes in Computer Science (LNCS)](vol. 3302); 2004
Most aspect-oriented languages provide only a fixed, built-in set of pointcut designators whose denotation is only described informally. As a consequence, these languages do not provide operations to manipulate or reason about pointcuts beyond weaving. In this paper, we investigate the usage of the functional query language XQuery for the specification of pointcuts. Due to its abstraction and module facilities XQuery enables powerful composition and reusability mechanisms for pointcuts.
Michael Eichberg and Mira Mezini and Thorsten Schäfer and Claus Beringer and Karl-Matthias Hamel
Proceedings of the 2004 Australian Software Engineering Conference; (ASWEC)
IEEE Computer Society; 2004
Policy enforcement is a mechanism for ensuring that system components follow certain programming practices, comply with specified rules, and meet certain assumptions. Unfortunately, the most common mechanisms used today for policy enforcement are documentation, training, and code reviews. The fundamental problem is that these mechanisms are expensive, time-consuming, and still error-prone. To cope with this problem, in this paper, we present IRC (Implementation Restriction Checker), an extensible framework for automatically enforcing system-wide policies or contracts. The framework is built on top of a platform for aspect-oriented programming at the level of Java bytecode instructions and is available as an Eclipse plug-in as well as a standalone application. It includes a set of directly usable checkers and can be easily extended to implement new ones.
Michael Eichberg and Andreas Sewe
BYTECODE 2010, 5TH Workshop on bytecode semantics, verification, analysis and transformation
Elsevier [Electronic Notes in Theoretical Computer Science (ENTCS)](vol. 264); 2011
New toolkits that read in, analyze, and transform Java Bytecode are regularly developed from scratch to get a representation suitable for a particular purposes. Nevertheless, the functionality implemented by the toolkits to read in class files and do basic control- and data-flow analyses is comparable and is implemented over and over again. Differences mainly manifest themselves in minor technical issues. To avoid the repetitive development of similar functionality, we have developed an XML-based language for specifying bytecode based instructions. Using this language, we have encoded the instruction set of the Java Class File Format such that it can directly be used, e.g., to generate the skeleton of bytecode-based tools. The XML format specifies (a) the format of the instructions and (b) the effect of each instruction on the stack and the local registers when the instruction is executed. This enables to generate generic control- and data-flow analyses such as an analysis that converts Java Bytecode into static single assignment form. To assess the usefulness of our approach, we have used the encoding of the Java Class Format to develop a framework to analyze and transform Java class files. The evaluation shows that using the specification significantly reduces the development effort when compared with manual development.
Michael Eichberg and Sebastian Kanthaka and Sven Kloppenburg and Mira Mezini and Tobias Schuh
Proceedings of the Sixth Workshop on Language Descriptions, Tools and Applications (LDTA)
Elsevier [Electronic Notes in Theoretical Computer Science (ENTCS)](vol. 164); 2006
Research related to alias protection and related concepts, such as, confined types and ownership types has a long tradition and is a promising concept for the design and implementation of more reliable and secure software. Unfortunately, the use of these concepts is not widespread as most implementations are proofs of concept and fall short with respect to the integration with standard software development tools and processes. In this paper, we discuss an implementation of confined types based on Java 5 annotations. The contribution of this paper is twofold: First, we discuss the incrementalization of the confined types analysis and second, we present the integration of the analysis into Eclipse using the static analysis platform Magellan.
Proceedings of Bytecode '05
Elsevier [Electronic Notes in Theoretical Computer Science (ENTCS)](vol. 141); 2005
The creation, transformation and analysis of bytecode is widespread. Nevertheless, several problems related to the reusability and comprehensibility of the results and tools exist. In particular, the results of tools for bytecode analysis are usually represented in proprietary tool dependent ways, which makes it hard to build more sophisticated analysis on top of the results generated by tools for lower-level analysis. Furthermore, intermediate results, such as e.g., the results of basic control flow and dataflow analysis, are usually not explicitly represented at all; though, required by many more sophisticated analysis. This lack of a common format, for the well structured representation of the (intermediate) results of code analysis, makes the creation of new tools or the integration of the results generated by different tools costly and ineffective.
Michael Eichberg and Mira Mezini
Proceedings of Software Engineering and Middleware: 4th International Workshop (SEM)
Springer [Lecture Notes in Computer Science (LNCS)](vol. 3437); 2004
In this paper we identify three problems with current component middleware. First, the implementation of services is usually not modularized, making it hard to adapt the platform to application specific needs, to exchange services to cope with changing requirements or to use it on different devices. Second, mapping components to objects results in a complex programming model and is making the component code dependent on the used component framework. Third, application level crosscutting concerns are not modularized. To solve these problems we propose an aspect-oriented programming approach complemented by standard Java 1.5 annotations to provide meta information about the components and a sophisticated query language for pointcut designation based on annotations.
6TH GPCE Young Researchers Workshop 2004
Journal of Object Technology; 2004
Middleware for component-based software development already provides some separation of concerns between the components implementing the business functionality and the component environment implementing the infrastructural services. However, the implementation of the services is usually not modularized, making it hard to adapt the platform to application specific needs, to exchange services to cope with changing requirements or to use it on different devices. Also, mapping components to objects results in code where the crosscutting concerns encapsulated in the middleware show up at several places, complicating the programming model and making the component code dependent on the used component framework. In this paper an approach to solve these problems based on the ideas of aspect-oriented programming is proposed.
Darmstadt University of Technology [D17]; 2007
Comprehensive tool support is essential to enable developers to cope with the complexity of modern software development projects. Software projects are getting larger and larger, are being developed using different languages, and make use of many third-party libraries as well as frameworks. Hence, tools are required: for software comprehension, for checking that libraries and frameworks are correctly used, and to ensure that the design does not degrade over time. Though numerous successful tools have already been developed for these tasks, several issues remain: the tools are usually highly specialized, their extensibility is limited, and an integration between the tools is lacking. Furthermore, IDE integration and in particular an integration with the incremental build process offered by modern IDEs is also often missing. Unfortunately, the direct integration of several code analysis tools with the incremental build process is not possible. When each tool processes the project’s resources on its own and also maintains its own source model, the overall memory requirements and analysis time is prohibitive. To address these issues, this thesis proposes the concept of a Build Process Integrated Open Static Analysis Platform. The core functionality of such a platform is to coordinate the execution of static analyses that are encapsulated into modules with well-defined interfaces. These interfaces specify what the analyses require and provide in terms of the data they process. For a tool that is built upon such a platform it is sufficient to specify the data it requires. The platform can then determine the set of analyses and their execution order to satisfy the tool’s requirements. Modeling analyses as modular producer-consumer units facilitates the simultaneous integration of several tools into the incremental build process of modern IDEs. When compared to using several independent tools, the overall memory requirements are reduced, since the source model derived by the executed analyses is shared among all tools built upon the platform. Furthermore, the overall analysis time is also reduced since analyses are executed at most once, even if the derived information is required by more than one tool. The overall analysis time is further minimized by the parallel execution of those analyses that process different information. The feasibility of the proposed approach is demonstrated by Magellan. Magellan is an open static analysis platform tightly integrated with the incremental build process of the Eclipse IDE. This integration turns Eclipse into an Integrated Development and Analysis Environment. The set of modules implementing the static analyses is freely extensible and the data model of the database is open. An open data model is crucial to support new analyses that need to store derived information for the use by subsequent analyses. Besides featuring a fully flexible analysis stack, Magellan also supports the embedding of query engines. Supporting the execution of queries is indispensable for enabling end-users to define application specific analyses. The ability to execute queries is also required to facilitate software comprehension tools. As a proof of concept an XQuery processor and a Prolog system are embedded into Magellan. Both engines are evaluated w.r.t. using them for the execution of queries along with the incremental build process. The XQuery engine is additionally evaluated in the context of software comprehension tools as a means to enable the end-user to define new ways to navigate through code. The platform is validated by four tools built on top of it: a software exploration tool, a metrics tool, an optional type system, and a set of lightweight static analyses that check structural properties of source code.
- Athene Preis für Gute Lehre 2012 (Hauptpreis) der Carlo und Karin Giersch-Stiftung an der der Technischen Universität Darmstadt
- Preis für die beste Lehre im Fachbereich Informatik an der TU Darmstadt im Sommersemester 2011 verliehen durch die Fachschaft Informatik
- Preis für die beste Lehre im Fachbereich Informatik an der TU Darmstadt im Sommersemester 2010 verliehen durch die Fachschaft Informatik