A Semantic Model of Reference Counting and its Abstraction (1986)
| Venue: | In Proceedings ACM Conference on Lisp and Functional Programming |
| Citations: | 64 - 0 self |
BibTeX
@INPROCEEDINGS{Hudak86asemantic,
author = {Paul Hudak},
title = {A Semantic Model of Reference Counting and its Abstraction},
booktitle = {In Proceedings ACM Conference on Lisp and Functional Programming},
year = {1986},
pages = {351--363}
}
Years of Citing Articles
Bookmark
 |
 |
 |
 |
 |
 |
OpenURL
Abstract
Interpretation of Reference Counting The semantics presented thus far is exact, and thus evaluating a particular reference count may not terminate, any more than a program in the standard semantics would. For use by a compiler we must choose a suitable abstraction (i.e., approximation) that will guarantee termination yet still provide useful information about the true reference counts. For us a suitable abstraction is one in which the inferred reference count is at least as great as the true one; i.e., we wish to err on the side of thinking there are more pointers to an object than there actually are. In this section we methodically develop such an abstraction in which: (1) the base domains are abstracted to powersets of finite approximations, (2) the primitive functions are abstracted similarly, (3) an abstract interpretation is thus induced on R p and R, and finally (4) a collecting interpretation of expressions is developed. 4.1 Preliminaries To set up things to come, we first gi...
