I can only imagine that these people complaining about everyone doing that are themselves the people who do that.

Why would you assume that? "He who smelt it dealt it"? As an observant driver, I notice this phenomenon because it creates a "channel" of other cars moving around me. This is hazardous because it involves a lot of unnecessary passing and lane-changing.

Probably also the same assholes who know there's going to be a long line for the ramp and, rather than getting into it and waiting their turn, stay a lane over and then try to merge in right at the end.

Well, thanks for the vote of confidence. This is something else that I see regularly at I-290 for the offramp to I-90/94. Actually, on this topic, I do have a pet peeve. When "merge ahead" signs are posted for construction, many drivers will merge shortly after the signs. However, if the signs are posted too far in advance of the actual merge, this results in some drivers merging too early, and other drivers seeing the opportunity to "jump the queue" and drive all the way until the merge point. If everyone stayed in their lanes until the merge point, there wouldn't be a moral hazard to jump the queue, and traffic flow would probably be improved due to an orderly "zipper" merge. But once an early merge has started, only an asshole will break the unspoken courtesy-merge and continue straight until the actual merge. All because the signs were so early.

Try it sometime.

I do, every day I drive in Chicago. "Every 5 seconds" is probably an overstatement, but this is a constant issue both on local roads and on the interstates/highways.

Wallace & Gromit

I haven't followed them recently, but is it fair to attribute this to Dreamworks? "The Wrong Trousers" is credited to Aardman Animations and the BBC; I think I discovered them through PBS. All that says about Dreamworks is that they have good taste in claymation inventors.

Star * movies were mostly popular in the western world, most people in the east dont give a damn what happens to these movies.

Be careful with that wildcarding! Are you positive that no Starcraft movie has ever been produced?

$16. Per seat. Plus the hideously overinflated styrofoam popcorn.

If you don't like the ticket price or the popcorn, why not just wait to rent or stream it at home?

that's what I meant

Yeah, I was hoping that. Sorry, you didn't really deserve a snarky reply, I was just titillated at the idea of a pure-userspace malloc under a multi-tasking OS. But seriously, see my other comment about glibc code. If a program does malloc/free above the size threshold for mmap, then every call does get passed on to the OS. Additionally, if your allocated footprint grows and shrinks by large amounts in LIFO order, you may see excessive sbrk calls. All of this is tunable by mallopt.

Calling into the kernel code every allocation / deallocation of dynamic memory is slow.

If you haven't already, I encourage you to download the glib source code and trawl around. Pick the latest code, or whatever you interact with on a daily basis. Check out the malloc code, it's not as daunting as it sounds. You should discover a few things:

• If your malloc size is big enough to trigger a mmap (128KB IIRC), then you are absolutely correct that the runtime library will invoke the OS on every allocation/deallocation. Until you hit your number-of-maps limit, at least. So yeah, repeated allocation and deallocation of large blocks of memory is bad in this implementation. But don't throw out the baby with the bathwater! mallopt can be used to tune the allocator to your program's expected behavior. As in: short-lived block sizes should probably be below the mmap threshold.
• If you don't trigger the mmap size threshold, then the allocator only goes to the OS if it needs to increase the size of the heap. Again, this behavior is tunable: how much to pull from the OS at once, and how much overhead to allow before returning space back to the OS.

If you're not writing software with high performance requirements, you can probably ignore all of this, because your program's bottlenecks are probably somewhere silly (and nobody cares, anyhow). If you are, this level of control is probably desirable whether you're using GC or deterministic deallocation.

True, but most of us aren't anti gun we are just pro gun regulation.

That's not an excuse for statistical lying. A solid pro-regulation argument should be able to address these nuances head-on and still be convincing. Instead, I keep hearing from people who are religiously pro-regulation or pro-gun, twisting the numbers to favor their hypotheses.

It's also scary ... how efficiently the authorities were able to put a stop to civilian movement *just by asking*.

This didn't come as a surprise to me, and so it doesn't scare me much. People in general respond to authority figures. I'm sure tensions were higher due to the bombing, and in a high-risk situation it is rational to comply with someone who is better-informed. I'm actually glad that this seems to have been a voluntary success; I haven't (yet!) heard any fallout from arrests, harassment, unlawful search/seizure, or excessive show of force.

To use the common analogy, I'm happy to be the sheep and let the sheep-dog take care of the wolf, since it's quite clear that the wolf is real (violent men with guns and explosives). It's only scary when the wolf is a figment created/fed/magnified by the sheep-dogs.

Thats why the suspect was found AFTER the lockdown, by a guy walking around outside his house.

That doesn't actually mean that the lockdown had no effect. At the end of the lockdown, more people than usual were in very familiar territory (home vs. work or transit); everyone was vigilant and focused. Had there not been a lockdown, maybe the "background noise" of daily comings and goings would have been enough to mask the guy's escape.

TOR and the idea behind TOR [is] probably the most important invention in internet and communication methodologies.

You don't think there are other more important Internet/communication inventions than Tor? Think lower... Hypertext, TCP, UDP, IP... Ethernet, Radio, Fiber-optics. Other stuff that is really cool, that I don't even know about. Tor is interesting in our times because privacy is currently struggling in the face of technology. But IMHO it's not nearly as earth-shattering as the stuff it's built upon.

What ELSE have they done to "impact technology"?

They've done a lot of little things for developers over the years. Some examples:

• Code Search: I used this to look for both users and implementations of open-source APIs. I haven't looked at it recently, but at the time it provided great results when searching for known functions. Definitely better than any full-web search.
• Utility libraries like sparsehash. Maybe the algorithm wasn't earth-shattering, but high-quality, independent, open-source implementations of data structures do have an impact on the global developer community.
• SPDY is Google's creation. My web browser, written by some other company, supports this protocol.