Google I/O 2009 day 1
I’m at Google I/O at the Moscone Center in downtown San Francisco, and today was the first day. Everything was bustling:
The opening keynote started with Google CEO Eric Schmidt, and I was worried wondering how he would make over an hour be interesting. He only took a few minutes, then Vic Gundotra, VP of Engineering, led the rest of the keynote which had many presenters showing off various projects, starting with 5 major HTML 5 features already supported in Chrome, Firefox, Safari, and Opera:
Matt Waddell talked about Canvas, the very nice drawing & animation API with pixel-level control. Brendan Gibson of Backcountry.com used this at SteepandCheap.com and sister sites for the cool People on Site graphs (with a workaround for Internet Explorer which doesn’t support Canvas yet). Also a quick demo of Bespin, an IDE in the browser.
Michael Abbott, SVP of Palm, showed off their webOS 1.0 which uses HTML 5.
A good summary of the 5 big features of HTML 5 is in Tim O’Reilly’s blog post about it.
DeWitt Clinton, Tech Lead at Google, showed Google Web Elements, embeddable Google apps similar to the way YouTube & AdSense have always worked. Currently conversations, maps, search. A blog post by Tim O’Reilly gives more details about Web Elements.
Romain Guy, Software Engineer at Google, showed off Android’s coming text to speech functionality. Then all attendees were told we’ll be receiving a new Google Ion (aka HTC Magic) phone, the unlocked developer edition, with a SIM card for T-Mobile giving 30 days of unlimited 3G data & domestic voice so we can play with it. That was enthusiastically received. Certain attendees such as myself were hoping there’d be a discounted way to buy one at the conference, so this surprise worked out nicely. :) Various people wrote this up in more detail. Here’s mine getting unpacked:
The rest of the conference was split into various tracks, and I stuck mostly with Google App Engine talks which were good. Most useful was Brett Slatkin’s on using Datastore’s list properties with separate entities just for lists that can be used just for their indexes in queries without serializing/deserializing the lists which avoids a lot of CPU overhead but is a little tricky to set up.
The after-hours party (dinner, music, silly video games, etc.) is now winding up, and a semi-drunk guy is walking around with a garbage can asking for laptops we want to throw away. I still need this one for a while longer, so I declined his helpful offer.