ylfink@water.waterloo.edu (ylfink) (04/04/88)
DEPARTMENT OF COMPUTER SCIENCE
UNIVERSITY OF WATERLOO
SEMINAR ACTIVITIES
RECRUITING/PROGRAMMING LANGUAGES AND
ARTIFICIAL INTELLIGENCE SEMINAR
- Friday, April 8, 1988
Dr. Stanley T. Shebs, of the University of Utah, will
speak on ``Automating the Wizard: Runtime Data
Structures For High-Level Languages''.
TIME: 11:30 AM
ROOM: MC 5158
ABSTRACT
Implementors of Lisp, Prolog, Smalltalk, and similar
languages have long faced a series of puzzling
questions about the runtime representation of data
objects (such as lists, symbols, and sets). How should
the type of an object be recorded? What is the best
memory layout for complex objects? Will multiple
representations offer any space or speed advantages?
How should storage be allocated and reclaimed? To
date, there has been little or no help for the
perplexed designer; the available literature has only
the sketchiest information about runtime data
representation, and even less about the relative
tradeoffs of different design decisions.
As a start towards remedying this situation, we will
look closely at the data structures in a number of
languages. From this past experience, we can derive a
number of rules and guidelines for implementors
designing new language systems. These rules are also
being encoded in an automated designer; its output is a
set of definitions for the compiler and runtime
components of Utah Common Lisp, a new and flexible
implementation presently under construction.