Pattern matching in an open world
Published: 7 April 2020
Abstract: Pattern matching is a pervasive and useful feature in functional programming. There have been many attempts to bring similar notions to Object-Oriented Programming (OOP) in the past. However, a key challenge in OOP is how pattern matching can coexist with the open nature of OOP data structures, while at the same time guaranteeing other desirable properties for pattern matching. This paper discusses several desirable properties for pattern matching in an OOP context and shows how existing approaches are lacking some of these properties. We argue that the traditional semantics of pattern matching, which is based on the order of patterns and adopted by many approaches, is in conflict with the openness of data structures. Therefore we suggest that a more restricted, top-level pattern matching model, where the order of patterns is irrelevant, is worthwhile considering in an OOP context. To compensate for the absence of ordered patterns we propose a complementary mechanism for case analysis with defaults, which can be used when nested and/or multiple case analysis is needed. To illustrate our points we develop Castor: a meta-programming library inScala that adopts both ideas. Castor generates code that uses type-safe extensible visitors, and largely removes boilerplate code typically associated with visitors. We illustrate the applicability of our approach with a case study modularizing the interpreters in the famous book ”Types and Programming Languages”.
Keywords: Meta-programming / Pattern matching / OOP
Scifeed alert for new publicationsNever miss any articles matching your research from any publisher
- Get alerts for new papers matching your research
- Find out the new papers from selected authors
- Updated daily for 49'000+ journals and 6000+ publishers
- Define your Scifeed now
Click here to see the statistics on "ACM SIGPLAN Notices" .