devinj

No, they can run on quantum computers, but to properly simulate a quantum computer, you may need exponential space/time. Googling for supporting data brought up http://alumni.imsa.edu/~matth/quant/433/shor-par/node19.html which discusses simulating Shor's algorithm on a classical computer.

The reason for the exponential is that the bits are all in a superposition of all possible states (so for a 14-bit register, that's a superposition of 2 to the 14 states), which operations all act on. Or something like that. I'm not really that clear on the details. :(

Then why does Bill Gates live in a relatively high-tax nation like the US, instead of Qatar? I think that the super-rich are willing to ignore a few percentage points in their income, for the sake of other concerns. They're so rich, that they can afford to pay the tax required to stay with their family and friends.

Read the entire article.

Well, "is" in Python is a specific notion of object identity, and Python happens to sometimes decide it's not worth creating new objects. Not really confusing, it's just completely different from '==' (except by coincidence basically).

Nope; that's a common misconception.

   >>> "1"*1000 is "1"*1000
   False
   >>> 2**31 is 2**31
   False

The fact is that the only thing guaranteed by "is" is that if you have "a = b", then "a is b". Whether or not "new" objects are going to be identical to old ones is left undefined, and may change depending on type or value.

Right, the memcpy() API is leaky because it leaves some things unspecified. You can deduce the implementation by providing various inputs to it that have "undefined behavior". This is a common problem in C, and there are usually functions that avoid it (Linus recommends memmove).

I guess I'm not with the times anymore. Thanks for sharing.

Valve did that, the FBI was just the tool that was used for the job.

I'm struck by the similarity in some respects of the two cases. I would guess that the same people that got really upset about Valve doing that would be a bit saddened by this, too.

Most people dev off of trunk when they're starting out, actually. If you're doing that, 1.3 doesn't offer anything exciting.

Actually, it's statistically indistinguishable, according to Google.

Of course it shouldn't. No facts should be surprising, but they are. So what do you do about it? You either fix peoples' expectations to match reality, or you fix reality to match peoples' expectations.

Considering that the expectations here seem pretty reasonable, I'd personally rather the latter.

I don't think that it was Python had anything to do with it. It's sort of like me calling my new programming language made for sysadmin work and bioinformatics "Pearl". It's just a confusing name considering that there's another programming language that's used for the exact same things, and spelled and pronounced identically.

Now imagine if I said on my website, "By the way, feel free to call it Perl, just be aware that there's also another project that calls itself that". Now I'm giving permission for people to be even more confusing!

Isn't your reply there the only reply on that thread? And doesn't it just link here?

Are you a bot or something?

I would share my desktop. I figure that the idea that somebody could be watching would be very motivating. And since the chances of someone actually watching are so slim, I wouldn't feel too embarrassed.

Not really. Unfortunately, while the Nazi doctors did a lot of stuff, it was almost all useless to the scientific community because they didn't really apply the scientific method. There are only a few citable results, in particular the Dachau hypothermia experiments (for example), and even those are rarely used for obvious political concerns. So actually, the Nazis have not advanced modern medicine very much at all.

My friends in the natural language field tell me Python and NLTK are more common than Java. Then again, this is at a sort-of Python-centric university (Toronto).

GIL isn't really relevant for blocking IO-bound code, because the lock is released (and not contended) when blocking. For nonblocking single-threaded code such as Tornado's event loop, the GIL also doesn't matter. Not to mention that the real speed concerns had nothing to do with interpreted versus compiled, because Tornado very clearly held its own, and web.py+flup (etc.) have architectural issues with regard to performance, far beyond mere interpreter overhead.

> Is bcrypt not the best possible security, or reasonably close to it?

Reasonably close to it, yes, but scrypt is better. scrypt makes the KDF expensive not just in time, but in memory as well.

http://www.tarsnap.com/scrypt.html

"We estimate that on modern (2009) hardware, if 5 seconds are spent computing a derived key, the cost of a hardware brute-force attack against scrypt is roughly 4000 times greater than the cost of a similar attack against bcrypt (to find the same password), and 20000 times greater than a similar attack against PBKDF2."

The geometry section looks like something I could have passed after taking high school geometry. i.e. an American or Canadian 14-15 year-old could probably do it easily.

It doesn't.

How is saying that non-autistic people have autistic traits any different than saying non-sociopaths have sociopathic traits? You're being hypocritical.

Certainly there are traits that are well established as being "autistic" and "sociopathic". This is the former, not the latter -- the latter is characterized by a more predatory or ambitious nature, which this guy seems to lack, at least by his own descriptions. It's morelike he doesn't see the point in socialization. Sociopaths, on the contrary, are very social.

And, despite the fact that you keep claiming I did so, I did not call him autistic, and the first time you accused me of calling him autistic, I said that wasn't true. I guess I'd hope I wouldn't have to tell you things twice.

I didn't say he was autistic. I said the described behaviour is autistic. Which it is.

Sociopathic is what it isn't.

This is more autistic than sociopathic. It's quite possible that he won't burn out, simply because he really isn't interested in socialization.

Every installer I've used that had any user-friendly implicit settings, also had an "advanced" mode where you could override them, in addition to the button that just installed the software with default settings. Why can't it be put there?

Why is it that this post is not dead, but the author's description comment is?

Schizophrenia is, to my understanding, sort of a catch-all phrase for several different problems with similar symptoms but different effective treatments. Even in the study, they found only 49% of people with schizophrenia were infected.

I've given friends directions to my building by showing it to them on Street View. It isn't devoid of upsides.

Seeing as plenty of people often use a computer that much or longer on a regular basis (not counting breaks to eat or go to the bathroom, obviously), I'm not really convinced it's just the how long you use a monitor. I'd suggest talking to an ophthalmologist for real advice.

That said, I think the usual line is to take a ten minute break every hour. Should be good for your legs, too. But really, talk to an ophthalmologist.

Yeah, there's an EBNF syntax description. I must have missed the part where the GP thought that's all that was necessary for a "real spec". I guess I just disagree.

Java specifies the semantics of Java and the JVM remarkably fully. Python, not so much.

Java has a real spec. Python has documentation, but it isn't really sufficient for a spec, and for the most part alternative implementations copy down to the last implementation detail (gc details notwithstanding). A test suite is probably better than what it has.