Safe strict evaluation of redundancy-free programs from proofs

Simple item page
dc.contributor.authorKnight, Bren Ericen_US
dc.date.accessioned2024-08-14T18:31:43Z
dc.date.available2024-08-14T18:31:43Z
dc.date.copyright1994en_US
dc.date.issued1994
dc.degree.departmentDepartment of Computer Science
dc.degree.levelMaster of Science M.Sc.en
dc.description.abstractMany investigators have considered the possibility of extracting programs in a functional language directly from natural deduction proofs in intuitionistic logic. This approach delivers programs littered with redundant components and has led people to examine various ways of pruning these useless subprograms. Here we give an algorithm for automatically performing this pruning without requiring the user to manually annotate their whole proof. It then becomes possible to allow classical reasoning at points in the proof whose putative computational content is redundant to the extracted program. One particularly good place to take advantage of this new ability is in a rule for well-founded induction which allows us to extract explicitly recursive programs. Unfortunately, all this new flexibility leads to a complication which has not been mentioned in the literature: by pruning supposedly redundant parts of the extracted program we destroy the subject reduction and strong normalization properties that are touted as one of the main advantages of the programs-from-proofs approach. The main original result of this thesis is a propositional type system which inserts a relatively small number of delay and force operators to restore these properties. We attempt to extend the result to a dependent type system based on first-order predicate logic via the erasure of dependent types.en
dc.format.extent156 pages
dc.identifier.urihttps://hdl.handle.net/1828/18462
dc.rightsAvailable to the World Wide Weben_US
dc.titleSafe strict evaluation of redundancy-free programs from proofsen_US
dc.typeThesisen_US

Files

Original bundle
Now showing 1 - 1 of 1
Thumbnail Image
Name:
KNIGHT_Brent_Eric_MSC_1994_677932.pdf
Size:
40.35 MB
Format:
Adobe Portable Document Format
Download

Collections

Electronic Theses and Dissertations (ETD)
 
University of Victoria Libraries

We acknowledge and respect the Lək̓ʷəŋən (Songhees and Esquimalt) Peoples on whose territory the university stands, and the Lək̓ʷəŋən and W̱SÁNEĆ Peoples whose historical relationships with the land continue to this day.