I was reading Andre’s running notes on GMail and thought I’d toss in a few bones.

leonard@gmail.com was taken and lhl@gmail.com wasn’t allowed. viagra@gmail.com was available, but I wussed out: randomfoo@gmail.com
HTML Tidy Online is useful for reading the code. Line/breaks and spacing are removed from the source. The content is also gzipped. I wonder how much space they actually save from white space compression. If they’re running through a compressor, why not better (single letter) variable names and such?
On page load, there’s an immediate check for document.all; also if you’re using IE, you need 5+ and XmlHttp support
OK, now we’re getting into the meat of it. Yes, there is namespace compression going on. There’s a frameset w/ an action frame and a display window. Within that are divs and iframes. I’ll do a full dissection when I actually have time. (Err that might not be for a while. Off the top of the head for the next two weeks: 2 papers, a class final project, 2 work projects, a AAA priority freelance project, MSR design project, getting car fixed, wrist looked at)
The suggestion that Google is using JS as ‘anti-spidering’ method seems pretty silly to me. Both Python and Perl have SpiderMonkey modules. That being said, obviously I’m really happy about a good DHTML interface for a webapp.

Jason’s been collecting GMail reviews if you want actually useful/insightful comments.