Microsoft Pivot: their notion of "very large data" is smaller than mine, but this is still a fascinating tool.
Via Jordi Cabot (again), a library of data models created by Barry Williams. Baseball umpire scheduling, genealogy, septic tanks, and many others, all in a uniform notation. I'll be using this in my next software engineering course.
Andrey Petrov, a former student and friend (sorry, that's ambiguous, we're still friends) is co-creator of Get Up and Move, which lets you barter exercise with friends. "I will do pushups for 1 minute if @shazow will dance for 2 songs." Still in very early stages, but a cool idea.
Google Summer of Code and UCOSP have both shown that it's easier for students to get into open source projects if there's a pile of tiny tickets for them to start with. The goal of the new OpenHatch project is to collect those together by searching the raffles  of various projects for tickets with appropriate tags.
Seems that a sys admin in Arizona installed SETI@Home on every computer in his school district. The result: 575 million hours of computer time over 9 years, costing $1.2-1.6 million in extra power consumption.