Meeting Agenda

Wednesday, June 5th, 1996

Location:: DC 1304
Time:: 1:30 PM
Chair:: Stewart Chao

1. Adoption of the Agenda - additions or deletions

2. Coffee Hour

Coffee hour this week:: Any volunteers?
Coffee hour next week:: Any volunteers?

3. Next meeting

Date:: June 12, 1996
Location:: DC 1304
Time:: 1:30 PM
Chair:: Wilking Chau
Technical presentation:: Richard Bartels

4. Forthcoming

Chairs:

Bill Cowan
Matthew Davidchuk
Ed Dengler
Saar Friedman

Tech Presenters:

Leith Chan
Josh Cameron
Stewart Chao
Wilkin Chau

5. Technical Presentation

Presenter:: Balasingham Balakumaran
Title:: N-Step Incremental Straight Line Algorithm

Abstract:

The problem of digitizing a line into a uniform grid has many solutions. Bresenham's algorithm uses only integer arithmetic and generates one pixel per inner loop. There are many acceleration techniques for it. The straight line algorithm described here is based on Bresenham's single-step algorithm but it improves pixel-generation efficiency by generating four pixels per loop. Bresenham's straight-line algorithm is the first in a family of N-step algorithms. In addition to illustrating the form of these algorithms for (N = 4), a technique for generating other members of the family is also described.

6. General Discussion Items

7. Action List

University of Toronto visit

8. Director's Meeting

9. Seminars

                   The University of Waterloo
                      200 University Avenue
                        Waterloo, Ontario

              The Institute for Computer Research

                      Presents a Seminar on
"Software Reliability Measurement Based on Processing of Failure Data"

by:     Dr. Karama Kanoun
of:     Laboratoire d'Analyse et d'Architecture des Systemes
        Centre Nationale de la Recherche Scientifique
        Toulouse, France

Date:   Thursday, June 13, 1996
Time:   1:30 p.m.
Place:  William G. Davis Computer Research Centre, Room 1304

Abstract:

Software reliability studies have long been limited to the appli-
cation  of reliability growth models to failure data collected on
the software under study.  The validity of the  results  obtained
is  usually  unpredictable:   they may be good for some models or
disappointing for all of them.  This unpredictability  has  urged
us to develop a method -- based on the processing of failure data
-- that helps reliability growth models give good  results  in  a
controlled  manner.   The  goal  of  the  talk is to present this
method, which is based on the combined use  of  descriptive  ana-
lyses,  trend  analyses and reliability models to control testing
activities, evaluate software reliability and  plan  maintenance.
Ultimately,  the  aim is, when used in real-time in parallel with
the development, to help manage the  development  activities  and
predict  software reliability.  The different steps involved will
be described and emphasis will be laid  on  their  aims  and  the
results  that  can  be  expected from each step.  This method has
successfully helped analyze the software reliability  of  several
real-life  systems.   However  there are still some limits in the
current state of art that have to be overcome to allow  efficient
application  of  software  evaluation  in the development.  These
limits will be discussed.  The application of the method present-
ed  to  a  real  software  system  will  show the adequacy of the
results obtained.

Everyone is welcome.  Refreshments served.



	**************************************************
	               Pure Math Applied Math 
	           Combinatorics & Optimization Club 
	                Summer Lecture Series
	***************************************************

		       	     TROY VASIGA

		  -----------------------------------
		      Cryptography: Then and Now
	    	  ------------------------------------
Abstract:
   Cryptography, the art and science of making and breaking secure 
   messages, has been part of humanity since the time of Caesar.
   With increasing information exchange on the Internet, and storage
   capabilities of "smart-cards" growing exponentially, the security
   of information is becoming as important as the information itself.
   This talk will briefly overview the history of cryptography and 
   as well as discuss the current state of the field, with many 
   practical and interesting examples discussed.

	           THURSDAY JUNE 6 4:30 PM MC 4040 
	            refreshments will be served.. 



CS Colloquium Series
Computer Science Department
University of Waterloo

Type Polymorphism for Object-Oriented and Distributed Programming

By: 	Dominic Duggan
Of:	Department of Computer Science
	University of Waterloo

Date:	Tuesday,  June 18, 1996
Time:	4:00 P.M.
Place: 	Davis Center, Room 1304

Abstract:

The last ten years have seen exciting work on new type systems for 
programming languages.  One of the key new concepts is "type 
polymorphism," and its combination with subtyping and inheritance in 
object-oriented languages.  Type polymorphism is currently the main 
consideration for extending the Java language and virtual machine.

I will describe some of the work I am doing in type polymorphism.  
This work includes a new static type system for object-oriented 
languages with type polymorphism.  This is the only type system to 
provide objects with polymorphic methods.  I will also describe an 
approach I have developed to providing type-safe user-definable 
marshalling for distributed polymorphic languages.  The approach is 
based on run-time types in polymorphic languages.

The target implementation language for these extensions is a new 
distributed programming language, designed as an extension to Standard 
ML.  Time permitting, I will discuss other work in this new language 
(including implementation).

No prior knowledge is assumed of type polymorphism, marshalling, ML or 
even Java.  A minimal knowledge of objects, as in Modula-3 or C++, is 
assumed.

Biography

A native of Dublin, Ireland, Dr Duggan obtained his PhD from the 
University of Maryland, College Park.  His accent appears to have been 
a casualty of this experience.  His research interests are in 
programming languages, programming environments and software 
engineering.  His work has appeared in several journals.  His most 
recent focus has been in the design and implementation of 
object-oriented and distributed programming languages, and in 
programming-in-the-large.  

Everyone is welcome