NuPRL

NuPRL
Name	NuPRL
Developer	Cornell University Computational Logic Group
Released	1980s
Latest release	ongoing
Programming language	ML
Operating system	Unix-like
Genre	Proof assistant, computerized proof system
License	Research

NuPRL NuPRL is an interactive proof development system originating from work at Cornell University by researchers in computational logic. The project influenced and was influenced by developments at institutions such as Carnegie Mellon University, Stanford University, Princeton University, and Microsoft Research, and it interacted with researchers associated with the Massachusetts Institute of Technology, University of Cambridge, and University of Oxford. NuPRL combined ideas from the work of logicians and computer scientists including Per Martin-Löf, Robin Milner, Dana Scott, Gordon Plotkin, John Reynolds, Robin Milner, Gérard Huet, and William T. Freeman.

History

NuPRL emerged in the 1980s from the Cornell Computational Logic Group under the direction of Robert L. Constable and collaborators including Stuart Berzins, Mark Bickford, and James McKinna. Its development paralleled projects at INRIA, the University of Edinburgh, and Xerox PARC, and it reacted to earlier systems such as AUTOMATH, LCF, and HOL. Funding and collaboration were provided by agencies and labs including the National Science Foundation, DARPA, the Office of Naval Research, and the Digital Equipment Corporation. Key milestones intersected with conferences and venues such as the Symposium on Logic in Computer Science, the International Conference on Functional Programming, the ACM SIGPLAN meetings, and the IEEE Symposium on Logic in Computer Science. NuPRL influenced subsequent efforts at the University of Illinois, the University of Washington, and the University of Pennsylvania, and it was discussed alongside systems from the University of Cambridge and Stanford such as Coq, Isabelle, Agda, and Lean.

Design and Architecture

NuPRL's architecture integrated a proof engine, a term language, and libraries managed by a user interface and batch tools developed at Cornell and contributors from Columbia University and Brown University. The system's design connected to concepts from the Lambda Calculus, Category Theory as used by Saunders Mac Lane and Samuel Eilenberg, and Computational Type Theory inspired by Per Martin-Löf and Henk Barendregt. Its layer model paralleled ideas from the Edinburgh LCF project by Robin Milner and Gordon Plotkin and incorporated influences from the ML language family created by Robin Milner and contributors at INRIA and University of Edinburgh. The implementation made use of tactics and decision procedures related to work by David Turner, Robin Milner, and Mike Gordon and leveraged user interfaces developed in collaboration with teams at Xerox PARC, IBM Research, and Microsoft Research.

Type Theory and Logic

NuPRL is grounded in a form of constructive type theory that reflects connections to the Curry–Howard correspondence, intuitionistic logic as articulated by L. E. J. Brouwer and Arend Heyting, and the realizability interpretations studied by Stephen Kleene and Per Martin-Löf. The system engaged with themes from proof theory explored by Gerhard Gentzen and Dag Prawitz, and with semantics influenced by Dana Scott and Christopher Strachey. NuPRL's type constructors and notion of computation relate to studies by Henk Barendregt, Gordon Plotkin, and John Reynolds on polymorphism and operational semantics, and the project interacted with categorical semantics from F. William Lawvere and Bill Mitchell.

Implementation and Tools

NuPRL's implementation used variants of ML and OCaml influenced by the ML family, and tooling incorporated contributions from teams at Sun Microsystems, Bell Labs, and Bellcore. Its development environment interfaced with editors such as Emacs, tools from the X Consortium, and windowing systems used at Carnegie Mellon and the University of Utah. Integration and interoperability efforts referenced standards and approaches from the World Wide Web Consortium, the Open Systems Interconnection efforts, and collaborations with projects at the University of California, Berkeley, and Los Alamos National Laboratory. The system's libraries and proof scripts echoed practices found in the Coq community at INRIA, the Isabelle community at Technical University of Munich, and the Agda community at Chalmers University.

Applications and Case Studies

NuPRL has been applied in formalizing mathematics and verifying software systems in contexts associated with the Department of Defense, the European Commission, and industrial partners including IBM, Intel, and Hewlett-Packard. Case studies drew on examples from programming language semantics pioneered by John McCarthy, Tony Hoare, and Robin Milner, and on verification efforts inspired by projects at NASA, Boeing, and Lockheed Martin. NuPRL was used in teaching and research collaborations with institutions such as Columbia University, the University of Toronto, the University of British Columbia, and the University of Michigan, and it contributed to formalizations comparable to work at the Institute for Advanced Study, Princeton, and Stanford.

Relation to Other Proof Assistants

NuPRL is often discussed alongside proof assistants including Coq, Isabelle/HOL, Lean, Agda, HOL Light, PVS, ACL2, and Twelf. Its constructive type-theoretic foundations contrast with set-theoretic and classical frameworks used in Mizar, Metamath, and Z3. NuPRL's tactic-based interaction shares lineage with LCF-style systems developed by Robin Milner and Mike Gordon, and its library and module practices intersect with efforts at INRIA, Microsoft Research, and Academia Sinica. Comparative studies referenced work by Gérard Huet, Thierry Coquand, Christine Paulin-Mohring, Andrei Voronkov, and John Harrison.

Reception and Impact

NuPRL has been cited in literature by researchers at Cornell, Carnegie Mellon, MIT, and the University of Cambridge, and its impact is visible in subsequent projects at Microsoft Research, INRIA, and Stanford University. The system informed debates and developments led by figures such as Per Martin-Löf, Robin Milner, John Reynolds, and Dana Scott, and it contributed to curricular offerings at universities including Cornell, Princeton, and the University of Pennsylvania. NuPRL's legacy appears in ongoing research at the intersection of logic and computer science involving conferences like POPL, ICFP, LICS, and CPP.

Category:Proof assistants