Heard strange noise from my car this morning, didn't realize what was wrong until I got on the freeway, and another drive kept horning/pointing at me. Pulled over, walked around the car, and the rear right tire was flat. Tried to go home with it, bailed out midway because the noise from the wheel really worried me. Since this was my first flat tire, I panicked a little, the fact that I didn't know how to replace a tire made it worse. The up side is I wasn't too far away from home, so I called up my brother and we switched cars. Arrived at work about an hour late.

After work, went out with a couple co-worker to 21st Amendment to ehh, celebrate the fact that I won't be bothering them during work anymore. Starting tomorrow I will be working from the Mountain View office. I didn't drink anything alcoholic since I don't want to risk getting a DUI. Stayed there for about an hour, went back to office, packed my computer, drove home. Got home around 9ish.

So the company I work for Riverbed Technology filed for IPO today. You can view their filing here. Google News has a couple results for the filing as well.

Totally unrelated, but NYT has a article titled Google in China: The Big Disconnect. At 10 pages, it will take more than a coffee break, but I honestly think it's a worthy read.

Been playing with oprofile at work lately. Not sure if it's the fault of the EMT64 intel CPU that's on my work computer or the fact that I am using a 2.6.9 kernel, neither CPU events nor callgraph work. I am not sure if I am allowed to discuss the profile results, but let's just say the result suprised me a little. Too bad sysprof requires kernel 2.6.11 or later, and too bad that Centos 4.2 is only up to 2.6.9.

Installed oprofile on my powerbook that's running linux 2.6.15, and I am still limited to using timer interrupts. In this setup oprofile doesn't complain about callgraph though, but I am not able to decipher the result that opreport -c produces. I used it to get a profile dump of gaim starting up, but the result isn't too useful, I am guessing that's because timer-based profiling doesn't really work well for short running tasks.

Next thing I am going to do is try oprofile on my home desktop, which has a AMD 32bit CPU. Will also try sysprof on my powerbook and see if that can produce anything more useful.

So I've been driving to San Francisco for quite a while now. It's not a pleasant commute, but it's not as bad as when I first started. I've been trying not to make the decision whether to stay up at SF or relocate to Mountain View. Mountain View is quite a bit closer to my home, but everyone on my team works at SF, and I am liking the environment there.

Well, the upside is, I don't have to decide now. The downside is, well, sooner or later I have to. I think I know what I want to do, but the ideal solution to the problem requires great courage to execute.

sigh