chancho

Don't miss the article on shitting dick nipples.

To paraphrase Kissinger, "FP language pissing contests are so vicious precisely because the stakes are so small."

(Also, why is this guy's blog written in first person plural?)

From that wikipedia page: "The untyped lambda calculus is Turing-complete, but many typed lambda calculi, including System F, are not."

Anyone have any details about System F not being turing complete? Is it not straightfoward to implement an untyped lambda calculus in a typed one?

But spending on r&d, in absolute terms, increased in the US as well as everywhere else, which is a win-win-win for everybody. It just happened to increase the most in China, but the US lead over China (why pick on China anyway?) is still huge, as the body of the article describes. Nice scaremongering headline, WSJ.

And at home you must be certain that you don't have anything non safe for wife.

Fuck Yes! quixotic relational databases (usually SQLite) EAV graph query metalanguage depressing W3C salted the earth RDF twin loadstones XML tree serialization SemWeb wankery SPARQL EAV graph databases NoSQL hype-wave AppEngine's BigTable-backed datastore exemplar APIs.

If I could send one message back 100 years in time, this would be it.

Yet he couldn't get concepts into upcoming standard. Far short of a dictator. There's also the significant fragmentation of the market (GCC vs. MSVC++) and the fact that both of these compilers pick and choose what to implement and when to do it (cf. partial template specialization.)

If Guido puts a feature in CPython, that's a iron-clad edict to all other would-be implementors, because CPython has a huge majority market share. Not implementing the same feature means you aren't implementing python. There is nothing equivalent in the C++ world so no one has power, either through standardization or implementation, to single-handedly steer the language.

" Let [Country X] achieve a certain status, then we can talk about relaxing the system."

Who is to say that any country has reached such an enlightened state. Your argument stands against any instance of democracy.

And a Protoss probe factory.

And a place Malcolm McDowell really wants to get to.

Computer Person.

I got my PhD without one dime of debt, and I think most of my classmates did as well. There is a lot of funding for grad students if you are flexible about what you want to work on. And if you absolutely MUST write your thesis on optimizing brainfuck compilers then you can still cover your tuition by working as a TA.

But this may just be the situation at my alma mater, a land grant university that attracts lots of research dollars but doesn't quite have the prestige to attract lots of talented students. I wonder if students at "name brand" schools don't end up with more debt because competition for funding is more intense.

I'm just as upset with the state of math education in the US as anybody, but if UW is anything like my alma mater, then "Atmospheric Science 101" is a class that liberal arts majors take to satisfy their math & science general education requirements. A lot of these students (my wife, for example) have serious math anxiety. She could probably get a 70% on this test, even today, so many years after school, but throw it at her without warning on the first day of a non-math class and she'd literally cry.

Now there's a case to be made that a good high school math curriculum should not leave anyone anxious about math, but I still think it's a bit of a leap to take this test and extrapolate to all students.

It's very easy to criticize a paper's execution while at the same time supporting it's line of inquiry. Just bookend the review with stuff like "This is a very important topic that deserves rigorous scrutiny" and "I fear the authors' failure to address the points listed above weakens their conclusions." This kinds of reviews are actually the most incisive, because it means the reviewer cared enough to really reason through your arguments and find the flaws in them.

What are some kinds (examples?) of code that you find ICC to compile better than GCC? Like, what's a typical loop that ICC can vectorize but GCC can't? I always have the damnedest time pinpointing when and where these optimizations fire and I've pretty much given up on the compiler when it comes to them. Rather I just develop code as normal and then when it's done find the top 3 or 4 functions in gprof (or Shark or whatever) and vectorize those by hand. Either that or try every compiler you have available and pick the one that yields the best time, but in my experience it's not always Intel.

Just buy one. Ubuntu is very slim to begin with (~4gb) and it's very easy to configure the filesystem to keep core files on the SSD and other stuff on a magnetic disk, e.g., symlink /home/kristianp/Music to /media/bighdd/Music or even just mount the hdd as /home. (Do they still make laptops with two drive bays? Used to be common but now that I think about it I can't recall any recent models with this. Maybe the Taiwanese brands? Asus, MSI, etc?)

I have a 16gb drive in my wife's netbook that has plenty of room still (no big music collection, though.) Our home computer has an 80gb drive, 70gb for Win7 and 10gb for ubuntu. Just to give you an idea, next time you boot up Ubuntu (karmic) watch the splash animation, the one that says "ubuntu" with a pulsing white progress-bar like thing underneath. The animation only cycles once and then gnome is instantaneously loaded and sitting idle (autologin is on.) Rebooting between Win7 and Ubuntu is like switching between Firefox and IE on my old computer. If you can afford one now, just buy it. If you wait for the "sweet spot" between price/performance/capacity you'll probably be waiting for a while. They'll be a better value in the future, but that will always be a case and it doesn't mean they're not a good value now.

I don't know what the average parental leave duration is, but only 12 weeks are guaranteed and that is unpaid. Seems like kind of a short amount of time to make such a significant dent in your career earnings. How many promotions do you miss in 12 weeks?

On the other hand, 9% of households are single-parent and 80% of those are headed by the mother. (http://en.wikipedia.org/wiki/Single_parent) That's 18+ years of hours lost to taking care of sick kids, ferrying them to/from school and activities, not being able to work late, etc. Seems more significant to me.

-------

I don't know if you were actually referring to parental leave (i.e., paid time off) when you said "paternity leave" or if you were talking about a parent who actually quits his job for years at a time to raise children from birth until adolescence. This seems to be the most significant factor determining income disparity and even luxurious Scandinavian parental benefits don't completely neutralize the intense societal (and probably biological) factors that cause women to make this choice, as the article points out.

No, neither of us are lawyers (let's keep this http://mattmaroon.com/2009/05/01/hacker-news-disease/ in mind) but since you've cited case law at me, then who am I to ruin an IP law speculation-fest.

Perhaps "it's illegal" is too bold a statement, but I do think that if tptacek got sued by HBO for torrenting True Blood then a fair use defense would not hold up. That's just my not-a-lawyer opinion, and I don't know of any precedent for this fair-use-timeshifting-by-proxy defense (do you?) but hey we might actually see a case like this soon. Last season HBO was monitoring the True Blood torrents and sending out C&D's to the ISPs of all peers with US IP addresses, if the Pirate Bay comments were to be believed.

Anyway, here's my not-a-lawyer reasoning: copyright law grants all rights of copying and distribution to the originator of the work and these are only transferred to a second party by written contract, EXCEPT for fair use (and first sale and other exceptions like that.) According to Wikipedia, fair use is a nebulous thing and varies from case to case but from what I can see it only gives you the right to copy for personal use (Tivo good) and never distribution (torrents bad). In fact Joel Tenenbaum and Jaimie Thomas both tried and failed to use fair-use as a defense for filesharing. So in order for hypothetical HBO v. Tptacek to not be summarily ruled in favor of the plaintiff, tptacek would have to show that his subscription to HBO constituted a contract which granted him the right to acquire and distribute HBO recordings. AND his torrent peers would all have to be HBO subscribers too. I doubt you'll find any language like that on his cable bill.

But what do I know.

It sounds as though the issue is whether or not the site responds promptly to takedown requests. Any torrent site that did would not be popular for long. They probably don't bother sending those requests to Google because Google doesn't host the torrent files, but if someone did manage to find a way to disseminate torrents using _only_ Google then you can bet Google would get on that. They do it with YouTube, and fast too.

> I'm slightly queasy about calling "torrenting whole seasons worth of shows I paid for" "time-shifting";

That's because it's illegal. You haven't actually paid for the shows. You've paid for access to the stream of content, but that doesn't automatically grant you access to acquire that same content from a third party. It'd be the same as if you saw it in the theater and then said to yourself "hey I've already paid for this once so I'm just going to torrent the dvd rip."

I'm not judging. I do this too, even for shows I don't "pay for", but let's face facts: we're minor criminals. I'm not entirely comfortable with it either so I'm weaning myself off of it. Eventually, everything becomes available on netflix, so really it just takes patience. (Sometimes my patience is worn thin by netflix's fucked up queue-jumping practices, then I torrent. I'm weak.)

Because then you need a motor and a battery, adding weight and cost to an already heavy and expensive item.

So if a part on your bike breaks then you're gonna cast a new one out of raw metal? Machine it yourself? How is this different than any other part you have to replace?

It's an electric motor, a battery and a cellphone in a red plastic shell. Nothing to get religious about.

"You can use your phone to unlock and lock your bike, change gears and select how much the motor assists you."

I'm guessing that "changing gears" is something that people in Copenhagen do rarely, perhaps using a wrench, like when you first get the bike or hook up a trailer. I miss flat cities.

On the other hand, a voice control gear changing app might be fun. I picture a dude on an all-white bike (tires too) wearing a clear-vinyl trench coat, telling his bike to "gear UP!" "gear Down...". The future is now.

Thanks for sharing your account of student life in Taiwan. It's always nice to hear firsthand about life in other cultures so that we don't hang onto our misconceptions. On that note, let me suggest that your caricature of American student life is a little cartoonish. Yeah it's like that in a lot of schools, but if your son or daughter goes to one of those schools then you live in a shitty neighborhood and should move. You're probably right about the means, but the variance is much higher in the US and its very possible to find an academic environment where a student can excel without fear of getting beat up. (Also, the bit about not making it into college being worn as a "badge of honor" is just silly. Why would they even apply?) FWIW I went to public school in Northern California (the central valley even) and I definitely got positive feedback from my peers for doing well in school, because my peers also took advanced classes. Apart from schools with gang problems, I would characterize it like this: In the US it's your choice whether or not you want to succeed academically. Your peers generally don't care because there are a myriad of other reasons to like or dislike you (looks, sense of humor, clique membership, etc.) So yeah there is not the social pressure to succeed, but for the most part there is no negative pressure preventing success (again, except in very troubled schools.)

But Avatars shaped like ourselves are still a few decades off, and needlessly complicated to execute. Avatars of our cars have worked out well.

While I don't agree that the US is "full" by any means, population density is a horrible statistic for comparison. Check this out:

http://www.wolframalpha.com/input/?i=population+density+of+e...

http://www.wolframalpha.com/input/?i=population+density+of+e...

Now that's of course massaging the numbers. People/area still puts the EU ahead, ~300 vs. ~500 p/mi^2 (got this by putting "mean" in front of the above searches) but the Eastern US is not "through the floor" by comparison. It's roughly equivalent to France.

Now if you lump the whole Western US in there then yeah you drop that number into the basement (Alaska is bigger than Germany+France+Spain) but if I take a big chunk of Africa or Russia and call it part of the EU then I can do the same over there, which is more-or-less the same thing from a geological, hydrological and climatic perspective.

So you haven't seen the movie and you don't know much about Cameron, Arrington or his site, and of those four things the one you are most interested in learning more about is TechCrunch?

> and because it should make the use of massively parallel computations easier. I concede that there are currently few implementations of massively parallel, functional code (MapReduce and possibly Data Parallel Haskell come to mind)

Not only are there few, their numbers are growing much more slowly than traditional C/Fortan/MPI environments. The "massive" in massively parallel is getting more and more massive every year. Running computations at full speed at these scales, 100K+ cores in multiple levels of hierarchy (same die, same board, same blade, same switch, etc) is a hugely mundane architectural problem. Not only is there no architecture besides MPI which has nearly as much effort invested in this problem, none are currently making the investment, so none have a chance to catch up within the next decade. The only way you will see FP doing massive parallelism at speeds even close to current best is if they wrap the C MPI interface, and I have yet to see any functional message-passing APIs which aren't tarted-up imperative environments. (If anyone knows any please share.) At it's core, message passing programming involves managing data, deciding where it resides and where it needs to go, which is kind of antithetical to FP.

You mention MapReduce, which I think is more of a data-center thing subtly but fundamentally different from HPC, involving less computation and more I/O. FP has a good shot there (cf. Erlang) but not much chance at cracking the HPC market.

Jack van Wijk gave the keynote at IEEE visualization a few years back and spent an inordinate amount of time talking about how to pronounce his name, the best part of which was when he told how Greg Nielson remarked that his last name looked like a tensor.

> If you need to get somewhere fast, use rapid transit.

I have a hard time believing anyone in the US (except maybe New York) could say this with a straight face. Where do you live that rapid transit is so rapid?