Wednesday, September 19, 2007

Keystroke loggers, etc.

There are so much computer malware and security compromises out there that if there existed a new operating system and hardware platform built from the ground up with security foremost, and even if nothing old worked on it -- all software needs to be reimplemented, I would still use it. I'm imagining, for example, a different LIBC or set of system calls to the operating system.

Tuesday, September 18, 2007

Hydrogen bomb

Cain strikes down Abel and wipes out 25% of the human population. How has the destructiveness of mankind's weapons and armies progressed when scaled by the growth of our population? What has been the bloodiest war relative to population?

Monday, September 17, 2007

Comment function

flip.const is a comment function. The ignored first argument can describe the second.

Monday, September 10, 2007

Chair

No single chair is the most comfortable. Switch among several different ones for long sittings.

Collaborative reverse engineering

Collaborative reverse engineering of AACS and binary-only Windows video drivers, hosted on a Tor hidden service. Collaborative annotation of assembly code and cycle-by-cycle execution, debugger, and interpreter.

Saturday, September 01, 2007

3rd place

A single-elimination tournament will determine without doubt who is the best. We assume there is no element of chance involved. It has an advantage over round robin that collusion between the weaker players to fix games cannot prevent the number one player from winning the tournament (Fischer accused the Soviets of this.) Single-elimination will always have one clear winner.

A double-elimination tournament structured with a losers' bracket appears to prevent collusion (I have not proved it), and there will be uniquely one undefeated champion and one runner-up who has lost exactly one game. However, that one loss may not have been to the champion, so there is the confusing possibility that someone who ultimately places worse than second place has actually beaten the runner-up. Can a tournament be structured to avoid this possibility?

Continuing the pattern, a triple elimination tournament will place the top three. The number of games needed becomes very large, while we wish to avoid the same contestants playing against each other more than once.

Is there a better way to design a tournament? We wish to avoid collusion being able to affect the outcome and a clear set of the top few placings at the end.