Moose (software)

Moose (software)
Name	Moose
Title	Moose
Developer	Genomic and software communities
Released	2003
Programming language	Pharo, Smalltalk
Operating system	Cross-platform
Platform	Smalltalk virtual machine
Genre	Software analysis, reverse engineering
License	MIT

Moose (software) is an open-source platform for software and data analysis used for reverse engineering, software visualization, and software repository mining. It provides tools for model extraction, metric computation, and interactive visualization for projects ranging from legacy COBOL systems to contemporary GitHub repositories, and is often used by researchers associated with institutions such as University of Bern, University of Lugano, and industrial partners including Google and IBM. The project builds on Smalltalk technologies and is tightly connected to communities around Pharo and Squeak, while influencing research in software comprehension, maintenance, and empirical software engineering.

Overview

Moose is a meta-modeling and analysis environment that combines model engineering, metric calculation, and visualization to support software comprehension tasks in academic venues like ICSE, FSE, and ASE as well as industrial settings such as Nokia and Siemens. It leverages the Meta-Object Facility ideas in a Smalltalk context influenced by projects like Moose (software) pioneers and tools such as EMF and Rigi, enabling researchers from ETH Zurich, INRIA, and KU Leuven to perform large-scale empirical studies. Moose supports heterogeneous inputs including Java, C++, Python, and legacy languages encountered in NASA projects and is integrated into workflows involving Eclipse and IntelliJ IDEA ecosystems.

History and development

Originating in the early 2000s from research groups at institutions such as University of Bern and influenced by work at University of Lugano and University of Geneva, Moose evolved through collaborations with researchers who published at ICSE and ECOOP. Early development drew on Smalltalk environments like Squeak and later Pharo, with contributions from academic labs and companies including Google and IBM Research. The platform matured through releases that responded to demands from empirical software engineering projects tied to conferences like MSR and initiatives at organizations such as CEA and the European Commission.

Architecture and core components

Moose's architecture centers on a meta-model (the "FAMIX" model) influenced by modeling standards like UML and OMG specifications, with a persistence layer implemented in Smalltalk and integrations for external storage such as PostgreSQL or Neo4j. Core components include a model extractor pipeline inspired by tools like SrcML and ANTLR, metric engines comparable to CK metric suite implementations, and visualization widgets akin to those in Graphviz and D3.js for dependency graphs used in studies by teams from Imperial College London and TU Delft. The system runs on Smalltalk virtual machines (VMs) related to Pharo VM and supports interactive inspectors familiar to developers from MIT and Stanford research groups.

Features and functionality

Moose provides model extraction for languages such as Java, C#, C++, Python, and COBOL through parsers and adapters influenced by ANTLR grammars, offers metric computation inspired by classical metrics from Chidamber and Kemerer and visualizations like treemaps, graphs, and matrices used in publications at VisSOFT and IEEE VIS. It includes refactoring-aware analyses similar to those in Eclipse JDT and supports history analysis leveraging Git and SVN mining techniques developed in MSR workshops. Interactive exploration features allow researchers from University of Cambridge and industrial teams at Microsoft Research to correlate design patterns, code smells, and architectural erosion across releases.

Use cases and adoption

Moose is used in academic research for empirical studies at conferences such as ICSME and MSR, in industry for legacy modernization projects at enterprises like Siemens and Nokia, and in governmental projects for large-scale code audits for organizations such as European Space Agency and NASA. Case studies have analyzed systems from vendors like Oracle and SAP and have been cited in theses from universities including University of Toronto and EPFL. Adoption is strong in research groups focused on software comprehension, software architecture, and software evolution in institutions like Chalmers University of Technology and Maastricht University.

Integration and extensibility

Moose integrates with tooling ecosystems such as Eclipse, IntelliJ IDEA, and Pharo IDEs and can be extended via adapters for parsers (built with ANTLR), graph databases like Neo4j, and visualization libraries inspired by D3.js and Graphviz. Plugin patterns echo architectures from Eclipse Plug-in Development and extension mechanisms used by GWT and OSGi, allowing collaborators from Siemens and startups to add support for domain-specific languages and continuous integration pipelines involving Jenkins and Travis CI.

Licensing and community

Moose is distributed under permissive licensing similar to the MIT License and developed by an open community comprising contributors from universities such as University of Bern, companies like IBM, and research consortia funded by entities like the European Commission. The project’s governance and collaboration model follow open-source practices common to projects hosted by organizations such as The Apache Software Foundation and incorporates contributors who publish at venues including ICSE, FSE, and MSR.

Category:Software engineering tools Category:Reverse engineering tools Category:Smalltalk software