Zeitschriftenartikel
Modular Collaborative Program Analysis
Lade...
Volltext URI
Dokumententyp
Text/Journal Article
Zusatzinformation
Datum
2024
Autor:innen
Zeitschriftentitel
ISSN der Zeitschrift
Bandtitel
Verlag
Gesellschaft für Informatik e.V.
Zusammenfassung
Static analyses are essential to ensure the efficiency and security of software. They face challenges as we use ever more and ever more complex software. We address these challenges by enabling collaborative analyses composed of small, maintainable modules. In this thesis, we propose the blackboard analysis architecture that allows independent modules to collaborate using a central data store. This architecture is framework-independent, applicable to a broad range of static analyses regardless of their implementation paradigm, and allows for modular soundness proofs. Using four case studies and an extensive evaluation, we show how the blackboard analysis architecture allows improving the soundness, precision, and scalability of static analyses and fosters the exploration of trade-offs between these qualities.