[Prev][Next][Index][Thread]
A. Black on Object-oriented types at Harvard
Date: Wed, 18 Dec 91 15:00:26 EST
To: colloquium, cs-faculty, cs-gradstudents, cs-research, cs-staff
Subject: CRCT Colloquium Series Presents
Everything you always wanted to know about types for
Object-oriented Systems
(but were afraid to ask)
Andrew Black
DEC and Harvard University
Monday, January 6, 1992
4 PM
Aiken Computation Laboratory 101
(Tea at 3:30 pm Aiken Main Lobby)
Abstract
This talk aims to make accessible to practitioners some of the
fruits of recent theoretical work on classes, types, inheritance,
type generators and the relationships between them. Although the
theoretical research community is still far from a consensus on
many topics, on many others there has been a convergence of views
and a deepening of understanding.
Sadly, much of the more theoretical work on objects, particularly
that which relates to types, is accessible only to trained logicians.
This is unfortunate, because programmers, language designers and
compiler writers have much to gain from a deeper understanding of the
tools of their trade; in particular, they need to know which problems
are solved (to avoid solving them again) and which problems are not
(so that they can be avoided).
Based on the speaker's experiences with the design, implementation
and modeling of the Emerald programming language, this talk will
attempt to shed light on issues like:
* what makes an object different from a value
* what types are, and why they are good for you
* what subtyping is, and why it isn't enough
* why contravariance isn't an unnatural act
* why inheritance is a relationship between programs, not between
classes
* whether you should care that TYPE is not a type
* ``polymorphism polyglot'' and how to avoid it
* what is the type of NIL, and whether it has one.
Host: Professor Victor J. Milenkovic