nspiegelberg

1. I added an extra brace after FBFatal in a final revision :( I'll ask for a revision 2. FBFatal has the same semantics as assert(), so that's correct. 3. The 'rebind_symbols' line truncates and is missing a horizontal scroll. You can view the rest of it if you click drag.

There are some technical inaccuracies that make it hard to look past as a coder (mixing up POSIX and UNIX, talking about unexpected behavior as a bug). However, I think the author did an accurate job on the high-level tenor of the article and touched on a critical meta-point about this bug.

SPDY (pre-Apple's version) + Open Source SSL Library + Perf Mods

Although fixing a bug requires analyzing the offending code, we weren't able to effectively narrow down the area of inspection through manual means (diff analysis, stack trace, git bisect). We instead narrowed down the area of code using sandboxing and non-trivial conditional breakpointing.

To fix the SSL library, we first used dup() to properly refcount the FD and then did more long-term restructuring to properly couple the FD & SSL object lifetime later.

As someone who worked in a flyover state for 5 years and then moved to SF (then NY)... YES! The comment about the 20% discretionary income and resulting impact is also true.

"I have spent the best years of my life giving people the lighter pleasures, helping them have a good time, and all I get is abuse, the existence of a hunted man." - Al Capone

I think NSA leaders would do well to read Dale Carnegie again and remember that even violent monsters think they are victims.

FYI: If you fast-forward to 2013, Philip Su is now heading the Engineering office at Facebook London. He's a really nice guy and great culture fit for the company.

Source: I work at FB. Also, just use the Googles...

You can learn that the decision was very complex with a lot of benefits and detriments for both stores. In the end, we thought that HBase gave us more initial benefits, reduced time to market, and a community with similar long-term goals.

Facebook is rated higher for a number of reasons:

1) High impact. Most team sizes are 5 people or less. Those 5 people are handling big systems like photos or memcache. You know you're doing important things.

2) Move fast. Example: Facebook Groups went from concept to release in 3 months. In other companies, it would take 3 months just to get through release engineering.

3) Hackathon! People here love to code, and the environment encourages you. A lot of managers have switched back to SWEs, reinforcing that it's not demeaning to stay technical.

4) Still growing. It's true, new hotness normally trumps old & established. However, where Google started doubling every 6-9 months; Facebook tries to maintain top quality at the occasional expense of slower growth. This is helping the company retain most of its original culture.

I am currently working on the HBase project at Facebook. It's extremely fun! We try to sync up with the OS branch on a weekly basis. Additionally, FB actually has an Open Source team whose primary job is to find out how we can use current open source solutions or remove company-confidential material and open source more of our infrastructure. It gives you real pride to know that your company completely supports your desire to help the software developer community.

Despite the facade of University Avenue, Palo Alto is a pretty tame suburban atmosphere. I tend to drive or Caltrain out to San Fran if I want to do something on the weekend.

Then again, if you have someone to hang with, all these awesome Open Space Preserves makes for great hiking/biking. My favorite is Astradero just because its a 5 min drive from my house.

Where is that open-source MapReduce stack? Facebook even gives their HPHP, Hadoop, and other code back to the community for free. The fact is, Facebook cares about Data & Google cares about technology. Google has just perfected making other companies look bad, while Facebook has the attitude of "let's keep working on our features".