Class 29: Some Reliability Considerations

Held Monday, March 13, 2000

Overview

Today we consider some reasons that programs fail, and what we might do about it.

Question 29 for today's class: What categories of people might be to blame when a program fails, and in what way might they have contributed to the failure?

Question 30 for Monday's class: What do we mean when we say a program is "correct"?

Notes

• A risks digest (20.83) is now available. Comments due Friday.
• I managed to snag copies of Database Nation, the extended Cathedral and the Bazaar, and an anthology on Open Sources from O'Reilly and Associates (thanks, ORA!). If you'd like to borrow one of them for break, let me know.
• For tomorrow, read chapter 10 of Dewdney (on Program Correctness)
• Just a reminder: I'll have sample stories ready after break. If I don't, I'll drop the requirement.

Contents

• A Taxonomy of Targets

Summary

• How do programs and computers fail?
• Some technical background
• Should we blame anyone? If so, who?
• Other questions?

A Taxonomy of Targets

• Let's begin by seeing who we blame and why.
• I'll then add a techie's perspective
• That's all we got to today