Some of the researchers talked about their work towards the Oxygen goal. I only got pictures of a few.

Anant Agarwal talked about the Raw computer architecture, where many computers on one chip are wired dynamically:


Anant Agarwal presenting the Raw Architecture

One of his slides drew a laugh when he talked about throwing gates (like "AND" and "OR") at the problem, since we just heard a speech by Bill Gates...


One of Anant's slides

Hari Balakrishnan talked about the problems of making a more adaptive namespace:


Building adaptive networks

David Clark talked about the Post-PC days. He talked about how general computing years ago (developed with LCS help) was timesharing of single computers by many individuals, later general computing was 1-1 with PCs. The Post-PC world will have many computers per person, connecting themselves as needed.

Here is a picture of him pointing to a cell phone earphone/microphone connection. In the future, if you tell a computer that you want to make a call, it will search around to find something like an earphone, something like a microphone, something that can communicate, etc., and viola!, a phone:


David Clark on the self-organizing Post-PC world

Victor Zue gave a presentation with a demo. "The central idea is to convert the computer from an instrument one programs to a partner in conversation" through speech recognition and synthesis:


Victor Zue presenting


The bottom line of his talk

John Guttag talked about, and demonstrated, doing radio-based communications as much as possible in software:


John Guttag delivering his talk

Here is a slide showing a radio as components next to one done as code:


Radio in hardware vs. software

Here is a diagram showing how they do it (A-D converter early after a wideband IF, then lots of fast software). The trick, they say, is to design for the average case, not the worst case; then software can be quite good:


How they do it

Here is a screen shot from a demo machine looking at a radio spectrum. Click on an area, and tell it to "listen" to the audio, and software tunes it in, doing all the demodulating, etc., for FM, AM, or whatever:


Software radio screenshot

Since it's all in software, you can even do TV. They think eventually they'll even do HDTV. Here is a picture of today's unit:


TV done in software

Unfortunately, these are all I have pictures from. I didn't get to hear them all.

Back to the Introduction