Please create an account to participate in the Slashdot moderation system

 



Forgot your password?
typodupeerror
×

Comment Re:Not just version 4.9 (Score 1) 739

But it's not. The language specification sets semi-deterministic rules for how the code must be converted to machine code. In this case the compiler was overstepping those bounds and generating machine code inconsistent with the source code and translation rules. Very inconsistent actually, given that the compiler was also ignoring the flags that imposed more stringent requirements on the translation than those in the language specification itself.

Comment Re:Or upgrade to llvm ... (Score 1) 739

Only for the 1% of your code that is the largest bottleneck, and even then a good modern compiler fed high-level source written with compiler optimizations in mind can often generate machine code that's as good or better than competent hand-coded ASM, due in part to the fact that it need not care about readability, maintainability, etc. Meanwhile assembly requires considerably more developer time to maintain than high-level source.

Good rule of thumb: stay away from assembly unless you've already:
- had a profiler pinpoint the code as a bottleneck area
- written the code with an eye towards being particularly compiler-optimzation friendly
- analyzed the compiler output and found significant bottlenecks that could be optimized away by hand, and aren't caused by sub-par input source
It's a pretty rare chunk of code that can make it past all three filters to be potentially worth rewriting in assembly. And those filters don't even consider the significant maintainability losses you typically get by introducing assembly.

Comment Re: I know you're trying to be funny, but... (Score 1) 739

So, he should cater to the personalities of the people that might theoretically contribute in the future, rather than to the personalities of those who have proven themselves by contributing for years or decades? Doesn't sound particularly effective to me. Perhaps it might be worth considering if the community were suffering from a shortage of contributors, but as far as I'm aware that hasn't been the case in decades.

Meanwhile, my sense is that Linus mostly only pulls out the abusive language in the face of gross incompetence - such as situations like this where major bugs make it into the release versions of critical infrastructure.

Comment Re:Meta-problem (Score 1) 512

Your government (I assume you are American) does provide foreign aid to Israel. It also supplies money and/or arms to a lot more unsavory countries like Egypt and Saudi Arabia. Sure, feel free to criticize Israel, but don't be hypocritical about it.

Can you quote me the part where I was being hypocritical? Can you show me where I said that I supported our funding of Egypt or Saudi Arabia? Or are you just trying to falsely discredit me because you don't like my opinion?

Comment Re:Meta-problem (Score 1) 512

Israel's not very efficient at committing genocide. Boko Haram in Nigeria has killed far more people. ISIS in Syria too. Etc.

The Israeli-Palestinian conflict is a minor regional territorial conflict. But because of huge anti-Israel sentiment among UN members, combined with a healthy does of Islamic racism, this minor conflict that could have been settled years ago is kept festering because the Islamic bloc at the UN sees it as a useful tool to weaken Israel. It's pure cynical geopolitics fueled by Islamic fascism.

Not for me. For me, the problem is that my government arms Israel. I accept that many nations handle their regional bood-debt feuds with more bloodshed. It's stupid and self-propagating, but fine, go ahead, if that's your thing. But if you blow my paycheck on your cock-waving bullets, you become subject to my judgment of your actions.

Comment Re:no problem (Score 1) 342

round != perfectly circular, and the Earth is a hell of a lot rounder than any circle you can draw by hand.

And if you're a pedant going with the mathematical definition of circle, then there isn't much at all in the universe that qualifies. Planetary orbits are probably about the closest approximations there are to a "perfect" ellipse, and the gravitational pertubations of other planets make sure they're far from perfect.

Submission + - Ask Slashdot: After TrueCrypt (slashdot.org)

TechForensics writes: (Resubmitted because was not identified as "Ask Slashdot"

We all know the TrueCrypt story-- a fine, effective encryption program beginning to achieve wide use. When you see how the national security agency modified this tool so they could easily overcome it, you'll probably understand why they don't complain about PGP anymore. The slip that showed what was happening was the information that NSA "were really ticked about TrueCrypt" either because they couldn't circumvent it or found it too difficult. From the standpoint of privacy advocates, NSA's dislike for TrueCrypt was evidence it was effective.

Next, NSA directly wrapped up the makers of TrueCrypt in legal webs that made them insert an NSA backdoor and forbade them from revealing it was there. It's only because of the cleverness of the TrueCrypt makers the world was able to determine for itself that TrueCrypt was now compromised. (Among other things, though formerly staunch privacy advocates, the makers discontinued development of TrueCrypt and recommended something like Microsoft Bitlocker, which no one with any sense believes could be NSA – hostile. It then became logically defensible, since NSA was not complaining about PGP or other encryption programs, to posit they had already been compromised.

This is the situation we have: all of the main are important encryption programs are compromised at least in use against the federal government. Whether NSA tools are made available to local law enforcement is not known. This all begs the question:

Does the public now have *any* encryption that works? Even if we can see the source code of the encryption algorithm the source code of the program employing that algorithm must be considered false. (TrueCrypt was the only program NSA complained about.) In the case of other software, it becomes believable the NSA has allowed to be published only source code that hides their changes, and the only way around that may be to check and compile the published code yourself. Half the public probably doesn't bother.

Okay, Slashdot, what do you think? Where do we stand? And what ought we to do about it?We all know the TrueCrypt story-- a fine, effective encryption program beginning to achieve wide use. When you see how the national security agency modified this tool so they could easily overcome it, you'll probably understand why they don't complain about PGP anymore. The slip that showed what was happening was the information that NSA "were really ticked about TrueCrypt" either because they couldn't circumvent it or found it too difficult. From the standpoint of privacy advocates, NSA's dislike for TrueCrypt was evidence it was effective.

Next, NSA directly wrapped up the makers of TrueCrypt in legal webs that made them insert an NSA backdoor and forbade them from revealing it was there. It's only because of the cleverness of the TrueCrypt makers the world was able to determine for itself that TrueCrypt was now compromised. (Among other things, though formerly staunch privacy advocates, the makers discontinued development of TrueCrypt and recommended something like Microsoft Bitlocker, which no one with any sense believes could be NSA–hostile. It then became logically defensible, since NSA was not complaining about PGP or other encryption programs, to posit they had already been vitiated.

This is the situation we have: all of the main or important encryption programs are compromised at least in use against the federal government. Whether NSA tools are made available to local law enforcement is not known. This all begs the question:

Does the public now have *any* encryption that works? Even if we can see the source code of the encryption algorithm the source code of the program employing that algorithm must be considered tainted. (TrueCrypt was the only program NSA complained about.) In the case of other software, it becomes believable the NSA has allowed to be published only source code that hides their changes, and the only way around that may be to check and compile the published code yourself. Half the public probably doesn't bother. (Would it not be possible for the NSA to create a second TrueCrypt that has the same hash value as the original?)

Okay, Slashdot, what do you think? Where do we stand? And what ought we to do about it?

Comment Re:Not news (Score 1) 342

The Americas were pretty ecologically lush when the Europeans first arrived, despite being pretty thoroughly settled for 5-10,000 years at that point. In fact the population was so dense that the journals of early European explorers report that the smoke from their cook fires was visible for a week before the land itself came into view. Now, a scarce 500 years later, the vast hardwood forests have been exterminated and the great plains have expanded from the Mississippi all the way to the Appalachians and are gradually turning into desert.

There's something to be said for actually paying attention to the ecosystem around you and considering the long-term consequences of your actions. from Wikipedia

"In every deliberation, we must consider the impact on the seventh generation... even if it requires having skin as thick as the bark of a pine." This is an often repeated saying, and most who use it claim that it comes from “The Constitution of the Iroquois Nations: The Great Binding Law.”

We might have been better off if we had incorporated a bit more of their constitution into our own.

Comment Take care of your yard? (Score 4, Interesting) 342

I imagine it's closer to "Invasive species are a danger to the entire ecosystem, including, eventually, themselves." When dealing with such the usual solutions are extermination (generally ineffective), or introducing a predator capable of keeping them in check without further destabilizing the ecosystem. Assuming we wish to do neither, nor suffer global ecosystem collapse, it would behoove us to start learning to co-exist with our ecosystem rather than strip-mining it.

And it's not like that is some sort of knee-jerk hippie "let's all live in mud huts" bullshit. As one example consider the gradually increasing numbers of oceanic "wildlife preserves" where all fishing and other destructive exploitation is banned - Not only does the protected area begin returning to pre-exploitation lushness, but so do the surrounding waters. Fishing yields around the protected zone reverse the global trend and begin to increase dramatically, greatly benefiting even the fishermen who were initially opposed to banning fishing in the richest waters. Given half a chance nature can be extremely bountiful, we just need to give the ecosystems a chance to stay healthy rather than maximizing short-term profits at the expense of long-term desertification.

Comment Re:no problem (Score 5, Interesting) 342

Very true. A few years ago I was tutoring at a community college and actually met a man who didn't realize the Earth went around the sun. At first I assumed he was pulling my leg, how could an American in this day an age not know that?!? But he was fascinated by the idea, and we had a long conversation about the basics of orbital mechanics and how they shape tides, the seasons, etc.

Slashdot Top Deals

Living on Earth may be expensive, but it includes an annual free trip around the Sun.

Working...