Meeting Agenda
Wednesday, 15 November 1995
- Location:
- DC 1304
- Time:
- 1:30 PM
- Chair:
- Richard Bartels
1. Adoption of the Agenda - additions or deletions
2. Coffee Hour
- Coffee hour this week:
- Ian Little
-
- Coffee hour next week:
- Julie and Andrew
3. Next week's meeting
- Date:
- 22 November 1995
- Location:
- DC 1304
- Time:
- 1:30 PM
- Chair:
- Ian Bell
- Technical presentation:
- Raymond Yiu
4. Forthcoming
- Chairs:
- Leith Chan
- Leo Chan
- Stewart Chao
- Wilkin Chau
- Tech Presenters:
- Richard Bartels
- John Beatty
- Ian Bell
- Leith Chan
5. Technical Presentation
- Presenter:
- Rob Kroeger
- Title:
- Safe Systems Programming
- Abstract:
-
A standard (Unix) scheme for systems safety involves
running system software in hardware protected memory,
with a fire wall separating user memory, and a program
written in an unsafe language. Any system
call is expensive, since it involves a copy of all data
from user to system memory, and a copy of all results
in the other direction.
A better approach involves no such fire-wall separation,
no hardware protection. It requires a safe programming
language. In this talk, safe means: static typing,
dynamic array bounds checking, control or elimination of
pointers, verified casts, and garbage collection.
With these elements strictly enforced by the compiler,
it becomes impossible to write systems software (e.g. a
device driver that must touch hardware registers), so
the language must provide a loophole mechanism (restricted
only to trusted users). An example of two languages
that provide some semblance of safety with loopholes
are: Modula 3 and Java. Modula 3 can have modules with
safe interfaces that encapsule unsafe inplementations.
Java allows C to be called and unsafe code can be written
in that language. Neither are completely satisfying
as safe system languages. The problem is still open.
6. General Discussion Items
7. Action List
-
Rob will co-ordinate the next UofT visit (they come here).
A proposal for February visit will be made.
8. Director's Meeting
There was none.
However, the directors would like to invite anyone interested
to attend the Fields Institute Workshop Thursday and Friday.
9. Seminars
The University of Waterloo
200 University Avenue
Waterloo, Ontario
The Institute for Computer Research (ICR)
and
The Centre for Advanced Studies in Finance
Present a Colloquium on
"Finance to Physics"
by: Dr. Joseph F. Traub
of: Department of Computer Science
Columbia University
Date: Thursday, November 30, 1995
Time: 3:30 pm.
Place: William G. Davis Computer Research Centre, Room 1302
Abstract:
The talk will consist of three parts. Some recent results and
open problems in information-based complexity will be described.
Some of the results will be applied to the valuation of financial
derivatives, an important topic in mathematical finance. The fi-
nal part of the talk considers whether it is possible to move
from limitative results in formal systems (Goedel, Turing, Cook,
...) to limits in empirical science. Are there provable limits
to what is knowable in physics and, more generally, in science?
_________________
Joseph Traub is the Edwin Howard Armstrong Professor of Computer
Science at Columbia University and on the external faculty of the
Santa Fe Institute. Previously, he was Head of the Computer Sci-
ence Department at Carnegie-Mellon University. He started
research in what is now called information-based complexity in
1959.
Everyone is welcome. Refreshments served.
The University of Waterloo
200 University Avenue
Waterloo, Ontario
The Institute for Computer Research (ICR)
Presents an Evening Lecture Series on
"Using the Internet as a Parallel Computer"
by: Dr. Kenneth Salem
of: Department of Computer Science
University of Waterloo
Date: Monday, November 27, 1995
Time: 8:00 pm.
Place: William G. Davis Computer Research Centre, Room 1302
Abstract:
The Internet is a vast and powerful collection of resources. It
serves as a communications system and as an information reposito-
ry. It can also be viewed as a computer consisting of millions
of interconnected, autonomous computers capable of operating in
parallel. The question is, can this computer, or a portion of
it, be programmed and used to solve real problems? Can its com-
putational power be tapped?
This talk presents some of the practical problems of computing on
the Internet. It describes some of the existing tools, such as
PVM and Condor, that support computing on clusters of intercon-
nected workstations. For a variety of reasons, such clusters
tend to be relatively small. It also describes the Distributed
Batch Controller (DBC), a tool designed to support data process-
ing on a larger scale. The DBC combines independent clusters
into a loose federation which can then be used as a powerful data
processing engine.
___________
Kenneth Salem is a member of the faculty of the University of
Waterloo in the Department of Computer Science. He joined the
department in 1994, after spending a year there as a visitor. He
has also been a member of the faculty of the University of Mary-
land in its Department of Computer Science, and a staff scientist
at CESDIS, NASA's Center of Excellence in Space Data and Informa-
tion Sciences. His research interests are in the areas of data
management and operating systems. Dr. Salem received his BSc in
electrical engineering and applied mathematics from Carnegie-
Mellon University in 1983, and his PhD in computer science from
Princeton University in 1989.
Everyone is welcome. Refreshments served.
10. Lab Cleanup (until 2:30 or 5 minutes)