4:15PM, Wednesday, April 15, 2008 <- Edit: I'll assume that's a typo, as tomorrow is the 15th and the correct date is listed here:

http://www.stanford.edu/class/ee380/Abstracts/

I think that's a typo, check http://www.stanford.edu/class/ee380/

Confirmed. I've gotten a few e-mails about this talk; it's definitely this year.

But will it have good support for w3c standards and a fast JS vm?

Frankly I don't care about its w3c standards and fast JS - those are problems people are working on, and it's not the interesting part about this browser. It's not meant to be a "production" browser that you will be using tomorrow to surf to facebook.

It's about applying the principles of operating system design to better use the browser as the next platform for applications to be built on. Google started this with chrome, and Gazelle takes it to the next level. Hopefully a major browser will utilize these principles for their next generation browsers.

Indeed. I saw a talk by Chris about Gazelle, and the numbers put it far behind IE8 in terms of performance (it is based on IE8, so that acts as a lower bound).

Starting from your own core, it might be possible to keep separation of concerns while not sacrificing performance so much. As an example, c.f. L4, which proved that microkernels are not necessarily slow.