Meeting Agenda - 2000.06.14

June 14th, 2000

Location:

DC1304

Time:

1:30 p.m.

Chair:

Daming Yao

1. Adoption of the Agenda - additions or deletions

2. Coffee Hour

Coffee hour this week:

Patrick Gilhuly

Coffee hour next week:

Teresa Ge

3. Next meeting

Date:

Wednesday, June 21st, 2000

Location:

DC1304

Time:

1:30 p.m.

Chair:

Mark Zadel

Technical presentation:

Mauro Steigleder

4. Forthcoming

Chair:

1. Shalini Aggarwal (June 28th)
2. Jason Ang (July 5th)
3. Ian Bell (July 12th)

Tech Presenters:

1. Daming Yao (June 28th)
2. Shalini Aggarwal (July 5th)
3. Mark Zadel (July 12th)

5. Technical Presentation

Presenter:

Selina Siu

Title:

Surface Reconstruction by Voronoi Filtering

Abstract:

Surface reconstruction from scattered sample points is a well known problem in computational geometry and computer graphics. Amenta and Bern published a new algorithm for reconstructing surfaces in 1998. I will present their algorithm and talk about my implementation and the results for my cs760m project.

6. General Discussion Items

• DC1304 is not available on June 28 and July 12. CGL meeting should be in Lab

7. Action List

• Grant applications to USENIX conference; see Robyn Landers rblanders@math, and http://www.math/~rblander/usenix.
  Deadlines: July 15, November 15
• SIGGRAPH early registration is next week (June 16th)

8. Director's Meeting

9. Seminars

• Wednesday, 14 June 2000, 3:30PM -- DC1304
  "Generic Algorithms"
  Algorithms and Complexity Group Seminar -- Computer Science
  John Lawrence, Dept. of Pure Mathematics, University of Waterloo

  A generic algorithm allows us to solve a problem for an algebraic structure independently of how the elements of that structure are encoded. We can often get lower bounds on the complexity of a generic algorithm by studying the solution sets of systems of equations. V. Shoup used this method to get lower bounds for a generic algorithm that solves the discrete logarithm problem for a finite cyclic group.

  In this talk we will discuss the nature of generic algorithms and give some examples of lower bounds for various problems.

• Friday, 16 June 2000, 10:30AM -- DC1304
  "Resolving Feature Interactions at Run-Time"
  Software Engineering Group, Computer Science
  Jonathan David Hay, PhD candidate, Dept. Comp. Sci., Univ. Waterloo

  One of the accepted techniques for dealing with large systems is to treat as separate concerns enhancements (called features) to the basic underlying service of a software system. These features may be created at different times (making the product extensible) and by different people (allowing features to be developed in parallel). However, features are not completely separate concerns because they modify the same basic underlying service. By doing so, features can conflict or interact in unpredictable ways. This is known as the "feature interaction problem."

  The problem is how to resolve this issue while still allowing features to be implemented separately. Some methods used to solve this problem is to design the system in such a way that feature interactions are guaranteed not to occur. Another idea is to allow feature interactions to occur but detect such occurrences and resolve them when they occur. The later is the focus of the presentation.

  The example system used in this presentation is the phone system. The phone system has a large number of features and is developed by many groups of people. This makes the phone system an ideal system to look at for solving feature interactions.

10. Lab Cleanup