Press Bio: Rick Rashid, SVP of Research, Microsoft

Editors note: Day 3 keynote started out a bit dry (overview of MS Research) and ended with the whiz-bang stuff. But as I baked on this, I thought that some of the points he made were still interesting/relevant (esp. the question "Why do basic research?"). So I've included both. Skip down to the "Whiz Bang!" section to hear more about the crazy fun stuff. Also, please forgive the roughness of the notes - I know some of the details are light here, but feel free to ask me more questions about this stuff.

Overview of MS Basic Research

• What's the Mission of Microsoft Research?
  • Expand the state of the art in each area we do research
  • Rapid transfer of innovative technology into production
  • Ensure Microsoft Products have a Future

• How is MS Research Organized?
  • University org model:
    • Flat structure, critical mass groups
  • Open Research environment:
    • Aggressive pub of results in peer-reviewed lit
    • Frequent visitors, daily seminars
  • Strong ties to Uni. Research: Nearly 15% of basic research budget directly invested into Universities worldwide
  • 850 PhD researchers across 6 campuses
  • List of many partnerships with universities and institutes around the world, but I noticed they seemed to not have any partnership with the University of Waterloo (I was curious about why - too bad huh?)

• What's the Value of Basic Research?
  • Source of IP and new product tech - 25% of all MS tech, more fundamental in nature
  • Problem solving
  • Early warning system

• Why do basic research?
  • Agility & ability to survive when things go bad
  • Having a basic infrastructure of smart people and wealth of technology to deal with and survive/change quickly
  • Vannevar Bush - Father of modern basic research justification (post WWII)

Whiz-Bang!

• Looking forward to 2020 - what is MS Research working on?
  • Provable systems?
  • SLAM - static driver verifier as part of the Vista DDK - used for analyzing large scale software components - allowed you to take orig program, things you wanted to prove expressed mathematically and conver into a new binary program of True/False --> Vista static driver verifier
  • Prove that programs terminate - lets us prove whether programs halts or not - "liveness" properties - eg every Acquire() must eventually be followed by a Release()
  • Concurrency analysis - "CHESS" ; code contracts; PEX - figure out what you should be testing; Z3 - ???; Proved Church's thesis (70 year open problem)
  • DryadLINKQ and Dryad - harness power of cluster computing - automatic query plan generation by DRYADLINQ and Automatic distributed exe by Dryad --> Executes across 1000s of nods, terB of Data
  • Presenter - Feng Zhao - energy use and the environment - Product: small battery-op wireless sensors and SensorMap software to aggregate and analyze data
    • Use of sensing, sensors for efficiency
    • Energy cost of applications - it used to be "how much mem does your program use?" - question is shifting to "how much power does your program use?"
    • Demo - sensors in a box
      • Sensor-Map - 90 sensors in the main hall
  • World Wide Telescope - virt. Observatory - 1.5MM active users - Sick!! 21 Giga Parsecs of visual information
  • Demo - Boku - lightweight programming language (for kids) - integrated gaming and coding on an XBox 360
    • Programming as a fundamental life skill
    • Demystify and engage
  • Demo: 2nd Light - Crazy! Goes beyond MS Surface - multiple surfaces - pre-distorted - crazy!

2nd year of Movember wraps up on a great note. Here’s the pic of my esteemed colleagues in the Mo-Down:

And for the 2nd year in a row, the winner for “Most Creative”…

Thanks to all who generously supported me and the other iMOson team members.