Results 1 
2 of
2
InductiveDataType Systems
, 2002
"... In a previous work ("Abstract Data Type Systems", TCS 173(2), 1997), the leI two authors presented a combined lmbined made of a (strongl normal3zG9 alrmal rewrite system and a typed #calA#Ik enriched by patternmatching definitions folnitio a certain format,calat the "General Schem ..."
Abstract

Cited by 821 (23 self)
 Add to MetaCart
In a previous work ("Abstract Data Type Systems", TCS 173(2), 1997), the leI two authors presented a combined lmbined made of a (strongl normal3zG9 alrmal rewrite system and a typed #calA#Ik enriched by patternmatching definitions folnitio a certain format,calat the "General Schema", whichgeneral39I theusual recursor definitions fornatural numbers and simil9 "basic inductive types". This combined lmbined was shown to bestrongl normalIk39f The purpose of this paper is toreformul33 and extend theGeneral Schema in order to make it easil extensibl3 to capture a more general cler of inductive types, cals, "strictly positive", and to ease the strong normalgAg9Ik proof of theresulGGg system. Thisresul provides a computation model for the combination of anal"DAfGI specification language based on abstract data types and of astrongl typed functional language with strictly positive inductive types.
Average Time Analyses Related to Logic Programming
 In Logic programming (Lisbon
, 1989
"... Logic programs are known to be amenable to parallelization. Our work is an attempt to quantify the magnitude of speedup one can expect from parallel execution of a logic program. To make average case analysis tractable we look separately at two aspects of logic program execution: the "subgoali ..."
Abstract

Cited by 5 (2 self)
 Add to MetaCart
(Show Context)
Logic programs are known to be amenable to parallelization. Our work is an attempt to quantify the magnitude of speedup one can expect from parallel execution of a logic program. To make average case analysis tractable we look separately at two aspects of logic program execution: the "subgoaling" aspect, which involves trying to prove a goal by using matching to reduce it to other goals and finally to true, and the "goal reduction" aspect, involving full unification. In the first case we assume that the andor tree determined by a goal can be constructed by matching only, and showusing the generating function approach of Flajolet that the average cost of evaluating such trees in parallel tends to a constant as the size tends to infinity, but that the same is true, though with a larger constant, for "clever" sequential evaluation. For the second aspect we use the generating function methods to obtain partial results about unification, which suggest that, although unification may...