The two-year storage is really the only part that bothers me. But the actual scanning doesn't, for some reason. I guess because people see my license plate every day anyway. It's a pretty public thing already, and it's government-issued so the only data being collected that they don't already have is my location, but again, any driver on the freeway can already see me. I don't know; usually I'm against most kinds of data harvesting, but for some reason this doesn't bug me as much. I guess because driving in your vehicle is such a publicly identifiable thing anyway, and it is on government property.

So Apple should pay hosting fees for apps not sold through their store? How does that make sense?

Just your friendly neighborhood reminder that Origin tracks your hardware, installed applications, software usage habits and more with no way to opt-out, unlike Steam. This is the new games industry.

Management doesn't know shit. Taking short breaks isn't slacking off, and studies have shown that such breaks improve worker productivity.

Management's problem is that it sees everything through a veil of pie charts and graphs, and if someone spends five minutes looking at pictures of their kids on Facebook, it must mean 0.2058% less revenue. Gotta fret over those graphs and spreadsheets.

Also, yeeeaah, can you come in on Sunday, too? We lost some people and need to catch up. Thaaaanks.

Yes, heaven forbid your employees take 10 minutes off from their monotonous cubicle hellholes to communicate a little with friends and family. It's not like studies have shown that more worker breaks increase productivity or anything. Henry Ford actually told his workers to work less because they got more done.

Well, of course they want participation. Participation means more information.

Google should focus on making their search engine better while thinking up the next big thing. Unfortunately, Google is so engineering-driven that it has a hard time understanding people. Even the use of "+1" comes off as mathematical and robotic. Grandma doesn't want to "+1 something".

Engineers often have trouble seeing their own work objectively, and they're afraid to apply human intuition in place of sampling data (e.g., the infamous 41 shades of blue). Google needs to change its culture so that it places greater emphasis on design and human interaction rather than technical impressiveness.

Neat, but is it surprising that transportation systems designed for the exact same purpose become mathematically similar over time? I'd be surprised if there wasn't emergent similarity in all urban transportation networks.

The engineer giving the talk mentioned that GCC eventually got precompiled headers over a decade after he had asked Stallman for the feature and after NeXTStep had already invested significant work in creating DevKit to work around GCC's limitations. It was those kinds of political moves that drove Apple to initiate the Clang project once LLVM came along.

It's not relevant how many use OS X for server tasks, because that isn't Apple's target market. Apple's XNU kernel is a mix of Mach and monolithic kernel features hand-tuned specifically for typical end-users. You can run certain kernel benchmarks and find that XNU is slower for certain things, but Apple isn't interested in working on them because it turns out they aren't useful benchmarks for the use cases they are aiming for, and I think that's a respectable position to take.

Besides, other factors contribute to performance bottlenecks, such as drivers. For example, Android isn't competitive with iOS when it comes to low-latency audio tasks even though Android is running on a server-class kernel like Linux and iOS is running XNU, a kernel whose performance you claim sucks. Like the other poster said, fretting over what kernel you're running just isn't all that relevant anymore given all the factors that ultimately contribute to a user's experience on a given platform.

Clang also addresses the lack of modularity in GCC. One of Apple's Clang engineers gave a talk several years ago about how at NeXTStep in the early 90s, he asked Stallman to implement precompiled headers in GCC. Stallman refused on political grounds. GCC was also intentionally obfuscated to make it difficult to integrate with. It's a clear example of rigid religious ideology standing in the way of and given prominence over technological progress.

After all these years of BSD code existing and thriving without issue, it's amazing that people still spread this kind of fearmongering despite the fact that this scenario has never come true.

The original code and its contributors don't magically disappear the moment a company makes a closed change. And if a company makes contributions it doesn't show anyone, you're free to make your own open contribution that competes with it. In fact, it's in company's best interests to rely on open contributions, because they don't want to waste time and manpower on, say, maintaining a compiler. This has proven to be the case with Clang. There hasn't been some evil proprietary fork that somehow ruined the world--and even if there was, people would just contribute free versions of the fork's features to the main tree. Companies are smart enough to know that this would happen and therefore realize that closed contributions of major features would be wasted effort.