ATerms are an efficient symbolic computation and exchange format. The core of the ATerm library is the C code that efficiently implements maximally shared (persistent, hash-consed) trees and lists, including an automatic generational garbage collector, and fast textual and binary serialization.
This library was developed at CWI with large contributions from INRIA Nancy and TU Eindhoven. It has been around for more than a decade. ATerm are used mostly in software products that heavily depend on symbolic evaluation, such as term rewriting (ASF+SDF, Stratego, ELAN, TOM) and model checking (mCRL2, haRVey).
Publications on ATerm library:
- Efficient Annotated Terms
- ATerms for the manipulation of structured data
- Generation of abstract programming interfaces from syntax definitions
- A generator of efficient strongly typed abstract syntax trees in Java
This aterms repository is a collection of components, all related to ATerms:
- apigen: generates C and Java code from ADT descriptions to wrap the lower level ATerm API in a typed abstraction
- aterm: code ATerm library in C
- aterm-csharp: experimental ATerm implementation in C#
- aterm-java: full ATerm implementation in Java
- aterm-xml: command-line tools for mapping between ATerms and XML
- balanced-binary-aterms: a balanced binary tree implementation on top of ATerms
- relational-aterms: an implementation of relational calculus on top of ATerms
- shared-objects: a core Java library for fast maximal sharing based on the factory and prototype design patterns
- shared-objects-csharp: shared-objects ported to C#