gpcz

Another great suite of debugging tools not mentioned is Valgrind ( http://valgrind.org/ ). The default application (Memcheck) can help find memory leaks and off-by-one errors. Cachegrind/callgrind are great supplements to Perf, and Massif can help get a better perspective on memory usage in your programs.

I feel reckless writing C/C++ code if I can't test it with Memcheck.

The Art of Unix Programming dedicates a chapter to the advantages of using text for the input and output of programs: http://www.catb.org/esr/writings/taoup/html/textualitychapte... . They also note that Unix and its derivatives are one of the oldest operating system paradigms still in active use, and that it has migrated to everything from smartphones to supercomputers. In a way, by using a Unix-like OS you are betting on text.

On the other hand, I would argue that engineering, architecture, and industry rely heavily on visual languages such as technical drawings, blueprints, control/process flow diagrams, etc. Text is a good bet, but it's inadequate for a lot of problems related to the physical world that require spatial reasoning.

The US Government had serious trouble reproducing FOGBANK ( https://en.wikipedia.org/wiki/FOGBANK ), and one cause was that even the original chemical engineers didn't fully understand the role of a certain impurity in a material. Is it possible to produce blueprints for a chemical plant that doesn't omit some inherent but important knowledge?

...then you end up in one of Peter Norvig's Google search history haikus: http://www.codersatwork.com/peter-norvig.html

Google and Facebook, which Gelernter says the very best people go to, have interview processes that filter out people who don't have a good command of fundamental topics taught in formal computer science college programs. Presumably the people who get into Google or Facebook are also passionate and self-motivated, most likely from an early age.

That implies to me that the self-taught+formally-trained people are in high demand and paid accordingly. Thus, Gelernter gets to choose among the people who are only self-taught or only formally-trained, and his sweeping viewpoints are based on his experiences with a subset of the workforce.

The "Relevant Conduct Provision" of the Federal Sentencing Guidelines allows uncharged but related criminal conduct to be used to determine your sentencing at a standard of the preponderance of the evidence. That means that if you're charged with a bunch of stuff and you plead guilty to one count, all the rest can be used to determine your sentence at a significantly lower standard of evidence.

So while the government didn't have to prove beyond a reasonable doubt that his tweets about funding were material support, those tweets could be used to increase his sentence for the Reza Niknejad thing.

I think you're underestimating the scale of this problem.

According to http://www.datacenterknowledge.com/archives/2013/01/18/faceb... , people upload 350 million new photos per day, and they have 240 billion photos total. If we use 1 MB per photo as an average, you'd need about 334 terabytes to store a day of Facebook's pictures. You'd need 228882 terabytes to store the whole thing. Assuming Wikipedia's estimate of $35/TB for desktop hard drives, it would cost about $8 million to store Facebook's photo archive (that's without videos or text). Actually hosting it would cost even more.

Getting in might be possible, but then the attacker would have to covertly exfiltrate and store all of Facebook's content to do a dump. Does J. Random Attacker have enough storage to do this?

Is an internationally-available Internet service running legally in its source jurisdiction obligated to prevent people from using it to break their local laws either legally or ethically?

The documentary "Once Upon Atari" explains the story of the consumer division (Atari 2600). Some of their best game developers left to form Activision after they tried to renegotiate their compensation with the new CEO and failed (this was post-Bushnell when Atari started making real serious money). After third-party games became legitimized, dozens of game companies sprouted up and tried to poach people from Atari. This put enough pressure on the Atari management to give release bonuses and $0.10/cartridge royalties that made a bunch of the developers millionaires.

Post-WW2 America had less income inequality than now, but it was far from a "very level and equal society."

In 1945, the share of the total income of the richest 1% was 11.07%, while in 2014 it was 17.85%. For the richest 10%, 32.64% in 1945 and 47.19% in 2014. A big chunk of the US segregated K12 schools and disenfranchised a race of people that politicians now use as a caricature in order to de-fund welfare programs. Women were kicked out of paying jobs after substantially contributing to the war effort because men came back and expected women to be homemakers.

The belief that post-WW2 America was some kind of glorious classless meritocracy is a myth. It just took a few decades of the natural self-reinforcing loops of capital to highlight the inequality that was already there.

If there isn't already an intricate system of landed class and privilege, who are "these folks over here" that give the money for the welfare and benefits?

At what scale would you consider it to be terrorism?

A related really cool thing I learned in grad school is that you can implement Deming regression (https://en.wikipedia.org/wiki/Deming_regression) by storing the moments (outer product sum and point sum) of the training points and then finding the dominant eigenvector of the outer product sum using a singular value decomposition. Since you can approximate the directional part of a 2x2 SVD with atan2(), it effectively becomes a O(1) operation to add or remove training points.

(specifics here: https://april.eecs.umich.edu/courses/eecs568_f12/linefitting... )

Using violent methods (such as intentionally sabotaging a car on a busy freeway with someone in it) to get media attention in order to further a political goal sounds a lot like the definition of terrorism.

Can you provide some links to the scientific papers asserting that media doesn't fundamentally change people's philosophies?

Also, if media can't change people's philosophies, what does according to science?

In this pizza recipe metaphor, are the recipes or the pizzas supposed to be the analogue to software?

If recipes, then the metaphor makes no sense. You can use the "recipe" that is LibreOffice to produce "pizza" that is work products that people would pay for (such as a lawyer drafting contracts, a consultant making spreadsheets, etc). Nothing in the GPL stops someone from profiting from the output of a GPL'd program.

If pizza, then I can only assume the "recipes" would be programs that produce programs, such as compilers or parser-generators. In that case the metaphor also falls apart, since GNU programs of that type have license exceptions that make the output of those programs (and the additional compilation files needed to run like libc and bison include files) usable without following the GPL.

In the early 90s during the era of game copiers, people made crack intros for Super Nintendo games and wrote SNES demos. I've been a fan of their music for years since zophar.net used to have an SPC collection (I can't find it on the zophar.net music page anymore, but you can find it from the archive.org version at https://web.archive.org/web/20070306011147/http://www.zophar... -- it's called "Public Domain SPC music"). An easier-to-access example is https://www.youtube.com/watch?v=pMOaN9SrR0M . As a lot of the work was done by pseudonymous people and it involved underground scenes, it's hard to find reliable history about these groups. However, there's enough homebrew SNES music from that era that I wonder if someone wrote a tracker.

If the group retains the attribution notices in the source code, hopefully this will be found when the VC does due diligence and then it's up to the VC what to do. If the group removed the attribution notices and you can somehow gather evidence of this, then they violated the Apache License and you should be able to sue. Whether or not this would be fruitful is up to whichever lawyer you contact.

Can you supply some details about the business-related dynamics at play here, specifically whether or not you're being paid any flat rate (or just equity), what your share is compared to the marketer, and whether you have a contract that puts you on the hook for anything?

A recent article ( http://www.hailpixel.com/articles/technical-cofounders ) and the accompanying HN discussion ( https://news.ycombinator.com/item?id=8959377 ) highlights how exploitative the "technical co-founder" relationship can be -- you may want to evaluate whether your situation resembles something like that and whether you want to re-negotiate or leave.

I did some searches for (all without the quotes) "privacy patriarchy", "privacy protects men", and "privacy feminism," but I mostly found old articles from the 80s and 90s about how Roe vs. Wade should not have been argued from a privacy standpoint and arguments about how drawing a line where family is "private life" can enable domestic violence and inequality due to the state being reluctant to interfere with people's private lives. Can you provide some examples from the last year of the trend you're describing?

This program makes me nervous because its primary use case is to obstruct justice, which the author tries to get away from with a thinly-veiled excuse that it could also be used to defend against criminals. I understand that technology is ethically neutral (for example, this program could be used to hinder reverse-engineering of a sealed computing appliance), but the fact that they're basically advertising this thing to obstruct police investigations puts me off.

The real meat of the defense-in-depth analysis is on page three of the article. Spyro had a two-layer defense-in-depth scheme: one layer that looked like a normal PSX cracking problem, and another that would look fine for a while and then mess up the game over time, which forced the crackers to make a complete play-through (and probably multiple failed play-throughs) to verify that their cracks worked. This served to make the cracker's feedback loops as long as possible. The author also acknowledges that it was impractical to add more layers of protection due to computational/IO/space costs, but that it would have offered more security, such as having multiple copies of the game's executable code on disc that are separately encrypted and randomly used, using custom compression algorithms, etc.

At its philosophical core, defense-in-depth is the idea of delaying an attacker rather than preventing an attack. In a military or IT situation this delay usually lets the defender detect the attack and counterattack/prosecute. In the cracking world, the delay IS the counterattack, since release groups measure their performance based on release quickness and the company (theoretically) gains revenue from the game not being on Kazaa during that critical sales season.

The biggest takeaway of this article is that effective security comes from proper threat modeling and analyzing the cost dynamics.

Most media companies in that era attempted to build an "uncrackable" system which always got cracked in short order because the mechanism depended on one tactic. By acknowledging that all protection schemes eventually get figured out and acknowledging the adversary's strengths and weaknesses, the author could then employ defense-in-depth techniques to maximize the cost of cracking the system.

Remember that every adversary has a budget.

I read this article a few years ago and thoroughly enjoyed it, but I've had trouble solving Default Permit in practice. Does anyone know of a free software operating system that white-lists its software (say by having the white-list be tied to the package manager) without having to do a lot of tinkering/configuring? I know you can do it with certain Linux add-ons that provide mandatory access control like SELinux and grsecurity, but it always seems like they introduce problems that you then have to troubleshoot for hours before your system is usable again. Bonus points for if this distro uses ASLR, W^X, and other countermeasures to avoid violating the security policy through code injection.

This phenomenon is a symptom of the Relevance Paradox [1]: people trying to solve a problem try to minimize how much information they get (seeing most of it as irrelevant), when a seemingly-inconsequential piece of information may be the solution. It's also a major factor of Hierarchical Incompetence [2] and part of why one really needs to think about the costs and benefits of establishing a hierarchy to achieve a goal. Hierarchies can be efficient for solving some problems and devastating for others.

[1]: https://en.wikipedia.org/wiki/Relevance_Paradox

[2]: https://en.wikipedia.org/wiki/Hierarchical_incompetence

I'm having trouble understanding the definitions of these roles. I see the chart, but the terms are all vague to me. What does a data scientist do that a mathematician or scientist doesn't do, and what does a scientific programmer do that a data scientist doesn't do?

My impression was that "data scientist" was a colloquialism for "statistician that knows how to program." Is a scientific programmer just a programmer that knows some statistics? Why is the direction important? The author says he/she feels that a programmer that knows statistics can make "more robust software" than the other way around, but what exactly does that mean? Do they mean "doesn't crash as much", or do they mean "gives the right answer more often?"

For the people asking questions about the experiment details, here is the full paper in Science (registration required): http://www.sciencemag.org/content/341/6143/275.full .

Based on the paper, it appears that they performed targeted mutations to manipulate the Mrap2 gene and fed the mice equal quantities of chow, and then they did the tests described in the parent article. At least to me, this is strong evidence that this gene affects obesity in mice.

Assuming that exposure is indeed the best way to treat PTSD, there is still the ethical matter of patient consent to consider. In the second linked article, the patient is exposed to the triggering content with full consent in a controlled clinical environment, but the recommendation in the first article is completely uncontrolled and could lead to adverse reactions away from a completely safe place. While I agree with Basoglu's assertion that there are too many possible triggers to warn everyone of everything, there are some common ones that we can warn about.

We'd still have the debate, but if we know in advance that we're going to show or describe the horrors of war in graphic detail, we'd warn the veterans in the audience beforehand. Trigger warnings aren't censorship, they're common courtesy.