Operational interpretations of linear logic

David N. Turner, Philip Wadler
1999 Theoretical Computer Science  
Two di erent operational interpretations of intuitionistic linear logic have been proposed in the literature. The simplest interpretation recomputes non-linear values every time they are required. It has good memory-management properties, but is often dismissed as being too ine cient. Alternatively, one can memoize the results of evaluating non-linear values. This avoids any recomputation, but has weaker memory-management properties. Using a novel combination of type-theoretic and operational
more » ... chniques we give a concise formal comparison of the two interpretations. Moreover, we show that there is a subset of linear logic where the two operational interpretations coincide. In this subset, which is su ciently expressive to encode call-by-value lambda-calculus, we can have the best of both worlds: a simple and e cient implementation, and good memory-management properties. (David N. Turner). 0304-3975/99/$ -see front matter c 1999 Elsevier Science B.V. All rights reserved. PII: S 0 3 0 4 -3 9 7 5 ( 9 9 ) 0 0 0 5 4 -7
doi:10.1016/s0304-3975(99)00054-7 fatcat:eglmzi4gkzhx7itc655phn3phq