But the idea is to be evil while looking innocent. Hard to read != innocent. None of the prize submissions uses easy to find complexity such as weird preprocessor stuff or spaghetti code. They all try to look as professional and straightforward as possible.
For instance, assembly code is right out - writing innocent-looking assembler is way too hard, and people will always suspect it.
Arrays, pointers and functions, no memory protection, dangerous strings. I would like to see the same contest with other 'safer' languages, say Java or Python.
What languages are best suited to underhanded tactics, that is, seemingly innocent but evil?. Notice that underhandedness is very different from plain old abuse -- anybody can write unreadable programs in their favorite language. But, can you make them "clearly read" something different from what is actually written?
Seems like an important question for people who use Open Source because of the difficulty for adding back doors. For many applications, security is at least as important as speed, and you already have The Shootout for that.
I am very interested in your obvious approach to solve shortages in food, potable water, basic shelter, education facilities and farmable land. Especially if it can solve everything at once, instead of focusing on small, well-defined problems.
Negroponte is not superman - but he has a valid point: kids/are/ the future. And to make the future better, you have to invest in education. If you can get kids to learn autonomously and to use computers - well, you're on your way to helping fix all that long list of shortages when the kids grow up.
The kids don't give a damn what the OS is. Hell, if you had never used a computer before, would you know anything about an 'OS' and its religious wars?
True, some people would like to promote OSS with the OLPC. It turns out that OSS is very customizable and has no licensing costs. So it is not such a dumb thing to do; and it got MS to contribute its apps at huge discounts just to not lose on the PR bandwagon. What's not to like? Affordable computer education for kids?
As long as the technology works and there is no recorded RMS greeting during startup... go Linux!
The/why/ is curiosity. Kids have lots of it, but you tend to lose it over time as they get slapped in
the hand and get told by adults to get serious. There's no telling to the number of great engineers (or doctors, or artists, or what-have-you) that we missed out due to stifled curiosity.
If you have a better way to build a mousetrap, build it and see if people will buy it. Trying to tell them they need it before you build one is... well, not how things work really.
The OLPC offers unlimited tinkering and very deep and broad educational (education as in building mental models of things and learning to learn, not as in rote memorization) experience for kids, and can help them learn to read and write and communicate and explore the 'net. It is not "a better mousetrap" - there was no mousetrap before, unless you are referring to the school itself as the mousetrap. And OLPC does not intend to displace schools.
Ok, the business model may not be too sound (but the entry of the ClassMate and 3$ Microsoft software bundles can be seen as partial successes - if the goal is affordable computing to 3rd world kids, things look much brighter than a few years ago). Yes, Negroponte is not a finance magician, and I guess he has learnt the hard way that large corporations do not always place developing nations before shareholder value - that's what PR is for, anyway.
Not to mention that there are simple combinations of moves that will guarantee that the cube will get solved. You can cram those in before an interview and voila! an expert 3D problem solver.
A better test for a programmer would be to say "write me a program that solves rubik's cube"... or if you want to test for 3D prowess,... graphically.
There will be mafias wherever there are groups or individuals with interests that run contrary to those of the state/society; you can't get rid of crime just by legalizing a few of the currently-illegal interests. Don't want to pay your taxes? Want to get your money back without having to sue the debtor and wait forever? Want something that the owner won't sell?
Not that legalizing marihuana would be negative, I'm just pointing out that crime is here to stay...
Are there any geeks who would not like to see wars turn into gigantic "battlefield" bot contests with armed bots shooting at other armed bots? For the sake of humanity I do hope that that is precisely what wars become. Or even taking it a step further and standardizing on some kind of networked multiplayer video game so that even machines need not be destroyed, just bits in memory.
Problem is, human troops will usually balk at shooting innocent civilians, or turning on their own country to impose a totalitarian state at the whim of a general. Robot troops would not give a damn, and would obey whoever controls the access codes. Be it president, general, evil genius, or rogue AI.
Yep, we aren't going to run short of apocalyptic science fiction anytime soon...
First, as rdebath argues, you only get 16 bits of CRC on TCP headers.
And furthermore, if you start calculating CRCs off random data, chances (>50%) are you will get a collision (two chunks of data with the same CRC) around the 256th try (this is known as the "birthday paradox" in criptography). Of course, to be really sure to get a collision you will need to try at most 65536 values; but you will reach a very high probability of clash much sooner than intuition may tell you.
Saying that the ceiling in neural nets has been reached just because no huge breakthroughs have occured lately ignores the fact that our own thinking procecess can very probably be modelled by a sufficiently complex neural net.
We only know how to teach current neural nets a few tricks. But saying that we have reached the end of the road is oversimplification. After all, understanding of thought procecess in the brain is still in its infancy. We don't know how our own "thinking machines" work (which are vastly superior to our current AIs), but we do know that they seem to be based on neurons dynamically signalling each other.
The core fallacy about it is that this doesn't mean crimes don't happen, it just means you won't hear about them. Which is, for the statistic, identical. It's a bit like closing your eyes and pretending that since you can't see the problem it doesn't exist.
It's not only used by children and politicians; there you have "don't shoot the messenger" (royalty and military), "the ostrich algorithm" (100% effective for technical woes), and "Gandalf Stormcrow" (LOTR version of the above).
> The only conclusion you can draw if you ask me is that it was obvious.
So, you're saying that because both Leibniz and Newton independently invented calculus, it must be obvious?
Thousands of college freshmen might disagree with you.
... but a few decades later, when math was better understood, it would have been obvious. Obviousnsess as in "apparent to most people with good knowledge of the area". If it's just one step further, its obvious. If it is requires a large leap, it is not. Shoulders of giants and all that.
I guess we will be needing lawyers on this one for the forseeable future.
Any system that allows a voter to prove to himself that his vote was correctly cast *after* it has been cast can be perverted into a vote-buying scheme.
I could place an stall with an internet connection outside the voting office with a big sign that read "20$ to anybody proving he voted $Party". Or I could be a bit less obvious, and just whisper around the neighborhood that bad things could happen if your slip didn't say what it was supposed to. You get the idea...
So in your qualified opinion, only people that already know unix/linux should use unix/linux.
I think this is great to allow new would-be linux users to install linux with less hassles. Or even experienced users to install linux withouth having to resort to burning a CD or downloading a bootable image into a thumbdrive. Repeat after me: convenience is good. Even if it allows people who shouldn't use it to shoot themselves in the foot. And I bet it can be easily adapted to boot any other distro, perhaps one of the more novice-friendly ones, such as Ubuntu. There's more than one way to use OSS. (Extolling the virtues of OSS on Slashdot: -1 Redundant, here I go:-).
By your line of reasoning, high-level languages should not exist, because people who don't know how to program assembler have no business mucking about with a computer, and dumbing it down is just going to make it easier for them to whine about it.
Germany has about 5 parties above the 2M voters watermark, many of which run local administrations. They can freely organize themselves into coalitions (that is, co-aligned parties can pool their votes together after an election) to achieve majorities when required. Voters have more choices, and there's not that many coalitions a party can enter unless it actually gets voted. Spain also has quite a few major parties. Don't want to vote for Socialist Party (social democrats in europe are like liberals in the US)? -- vote for the United Left party; they have similar goals and are likely to form a coalition if neither can rule by itself. Not that the system is perfect, but you do get slightly more "competition for ideas" and accountability than with a party system where minority parties don't get to play at all. There's lots of ways to organize a parliamentary democracy.
The actual study demonstrated that, for a given population (details on the selection of which were not provided), those that self-described themselves as liberals (whatever that label means) reacted differently to a situation (monitored by a certain measure of neural responses) than those that self-described themselves as conservatives. That part is "good science", though I may find it inconclusive. The bad part is where the study itself (from here) concludes that
Stronger conservatism (versus liberalism) was associated with less
neurocognitive sensitivity to response conicts. At the behavioral
level, conservatives were also more likely to make errors of commission.
This is only true within the experimental population, conservatism and liberalism where asked (instead of "measured"), and the neurocognitive sensitivity can depend on a host of other factors (mostly dependent on the selection of the experimental population), which may better explain the differences between subject's reactions. This part I find wrong.
When you read it, you find this:
a study that found a pattern correlating the way certain people behave and think There may be a pattern that correlates the way people think and their "left vs right" positioning -- but I will refuse to understand that "conservatives just can't help being hard of head" (tempting though it would be:-p) unless I see much stronger proof. Considering your opponents stupid is tempting, but leads to zealotry. Have I just self-defined as Liberal? Guess it's high time I had my brain scanned...
But the quoted excerpt from the article noted that this single-item measure accounted for a whopping 85% of the variance in Presidential voting intentions in surveys across three decades [...] I would much prefer to see something like "those who last voted republican" or "those who last voted democrat" instead of "conservative" and "liberal". It would be 100% acurate (assuming nobody lies), and you could then analyze the large percentage of people who don't bother to vote.
Also, your definition of "liberal" and mine are bound to be quite different; I happen to think it is a gross oversimplification, which fits nicely into a two-party system where many end up thinking there's only two possible stances on each subject (and furthermore, that the one espoused by their "team" is the right one). I think it is obvious that it would be more accurate to quiz people for key issues where "left" and "right" differ, and see where they actually end up when you map the answers to the scale. Or even better, try to correlate with particular stances on particular issues. Asking "conservative or liberal" when neither is clearly defined is, IMHO, flawed.
Your assumption of a.712 on some fictional absolute scale having to have the same stances as another who scored.712 is ridiculous.
I didn't say that. I used it as an example of what's wrong with the simplification. It's like using a "strawberry vs chocolate" metric to find out what you like to eat. Politics should be about taking stances on issues. Political marketing is about convincing you that it's all reducible to "us vs them", and that the only thing you should worry about is cheering your team or bashing the opponents.
It's a matter of how many of your views tend to be liberal or conservative. It's not like if you're a little conservative, you're going to have the same stance on a particular conservative issue as someone else who is a little conservative.
So you say there is a full "conservative" view and a full "liberal" view of the world, and that both of us can agree on what an ideal conservative and an ideal liberal would say on any topic?. That's plain wrong. Hard problems do not have unique "left" and "right" viewpoints.
[...] Political leaning is a part of a human behavior. Everyone leans in one way or the other. Why is it so crazy that those who tend to lean towards one direction tend to think a certain way?
Because I think that oversimplifications are dangerous; and I do not think that the paper supports this particular generalization. I also think that "liberal" and "conservative" are moving targets, and to that extent are more like marketing brands than real products.
[...] On an off-topic note, more than two parties is dangerous because you run the risk of the majority of people not voting for the winner. If you have candidates x, y, & z where y & z share a lot of similar views except one, you can get a situation where 3/5 of the population are split amongst those two. Now, 2/5 of the votes went to x, which 3/5(obviously) of the population did not vote for, but even so, he got more votes than the other two (the other two both getting 3/10 of the vote, whereas (if you can't do math) he got 4/10). That's why more than two political parties is illogical. You can get a winner who can possibly share no views with the majority of the population.
You are describing a "winner takes all" approach. AFAIK, that's very rare. In my country, x&y would have to join forces, or x&z or y&z. Government requires 50% +1 support to be elected. If no agreement is reached, voting is repeated. There's a whole theory on voting out there, and it *is* possible to get multi-party democracies off the ground. In fact, most of the democracies of the world work that way (at least nominally; many are de-facto 2-party systems). Start by checking wipedia.
The bad part about only 2 parties is that you either vote for one, for another, or abstain. You always end up voting "lesser evils", and the system lives on. If there were credible contenders, with which alliances could be formed, you could vote for co-alligned "y" and thus transmit a message to "x" that yes, you are mostly co-aligned, but somehow pissed at them. If there's only "x" and "z", you'd have to either abstain or vote "z". Abstaining and voting blank (marginally better) are too similar, since there are too many ways to interpret what your goal was when doing so. Two-party systems also promote finding where the "middle ground" is and appealing to it (since that is what decides an election if you manage to get everybody nicely layed out in a 1-d spectrum). That gives you reactive, poll-driven politicians which have no real opinions, instead of active politicians who can stand for the ideas you voted them for.
[...]You realize all this study is saying is that those who define themselves as liberals/conservative tend to share a similar way of thinking as others who label themselves as liberals/conserva
From the real article (not the newspaper writeup):
female)Subjects reported their political attitudes condentially on a -5 (extremely liberal) to +5
(extremely conservative) scale. So yes, they did fill it in themselves.
You say that "liberal" vs "conservative" is a useful spectrum characteristic. I say bollocks. They convey no information unless you define the extremes first, which (surprise) usually coincide with (possibly extrapolated) party lines. So a "liberal" vs "conservative" axis usually means "position yourself in a segment between these two parties". What do I stand for if I score 0.712 in an absolute scale from "conservative" (0) to "liberal" (1)?. What's a 0.5? Certainly not my position on privacy, copyright, patent reform or a dozen other topics, since those don't usually show up as party priorities.
No. As an older person who goes to college, I can tell you that college doesn't change political beliefs, nor are college students open to more ideas than the general population. Other studies have shown that college students are more likely to have a particular political affiliation, conservative or republican, than the general population.
Yes, but depending on the college, the prevailing political opinion may be heavily slanted towards one side. That would certainly skew the results, as people who couln't make their minds for themselves would be answering with locally "righteous" ideology, and cases of those who did not cave in would be more extreme (either because they felt strongly about their options, or because they stuck to their choice out of being stubborn). My wild guess is: predominantly liberal college, few conservatives to choose from, most happened to be headstrong.
Repeat the experiment with a different distribution to check for this bias, or quiz people on their political views instead of allowing them to tick a box.
People are not uniformly mixed. The next three persons to walk by my door are likely to be my colleagues returning from a coffee. How do you take a "random" sample of 43 from millions of non-uniformly mixed citizens? If you go through the pain of finding a really significant, random sample, would you only mention "right-handedness" and "sex" (65% female) as variables in your paper? And not (say) religion, or education level, or political makeup of place where the participants were found? All of these sound important to assess if a correlation of "conservative" or "liberal" to anything else within the sample population can be extrapolated to anything meaninful in the broader population.
Working in constant touch with researchers, "random populations" in academia are usually whoever happens to be working in or near the lab when the paper's deadline looms near; or students who happen to be in the classroom. Not "random" for any statistically sound definition of the term. For starters, I would have found it more significant if the researcher could specify the distribution of "liberal" vs. "conservative" in the population pool from which the participants where selected. There's places where it is easy to say you're republican, and others where it is easy to say you're liberal.
Also, I do not think that asking people their perceived "liberalness score" actually equates to people standing for the relevant values (and the same could be said of "conservativeness"); as in "I really think this is right" vs. "I say this because it is accepted party dogma".
The linked site is scarce on details - the paper itself is more interesting. First, David Amodio (lead researcher) is not obviously flaming. I'm no expert in neuroscience -- but the data looks good, and he has a track record on usage of scans for similar tasks (most of it race-bias related, but that's another subject).
Here's an interesting part of the experimental design:
To test the hypothesis that political liberalism (versus conservatism)
would be associated with greater conict-related ACC activity, we
recorded electroencephalographs from 43 right-handed subjects (63%
female) as they performed the Go/No-Go task. Subjects reported their
political attitudes condentially on a -5 (extremely liberal) to +5
(extremely conservative) scale. This single-item measure has been
found to account for approximately 85% of the statistical variance in
presidential voting intentions in American National Election studies
between 1972 and 2004 (ref. 8). Among participants in the present study
who reported voting in the 2004 presidential election, a more liberal
(versus conservative) ideological orientation strongly predicted voting
for John Kerry versus George Bush (r(21)= 0.79, P o 0.001).
I think that there are two ways in which the experiment may be flawed. One is that 43 persons are not enough to extrapolate to the whole US population; and more importantly, no details are given on how they were chosen. If they were chosen among colleagues in an academic setting, where most people (your mileage may vary) are left-wing, you may have problems finding people which self-describe as conservative. These few would be most resistant to changing their viewpoints, I would guess -- since otherwise they may have flipped from exposure to liberal arguments.
Another way in which I think the study may be flawed is by asking people to self-define their position in the political spectrum -- a one-dimensional political spectrum. What guarantees do you have that participants really are "conservative" or "liberal" (whatever that means to you), and have actually thought about the political issues involved in each "choice" (as if there weren't many, many greys)?. A 2-dimensional political positioning would provide more insight. A short questionnaire where participants actually had to think, instead of "choosing their favorite color" would have been even better.
This is assuming that the researcher knows what he's doing, and the
conflict-related ACC activity was indexed by two ERP components. ERPs are scalp-recorded voltage changes reflecting the
concerted firing of neurons in response to a psychological event. is actually a good measure of resistance to change or willingness to accomodate it. No details are provided on the exact activity, other than stating that parcicipants were offered the choice of two actions, "Go" and "No-Go".
Where the character played by Lukas Haas is originally a rabid republican, until the doctors find that he has a tumor in his head. Once the tumor is removed, he's instantly converted (back) into a left-wing liberal - just like his entire family.
Quotes:
(father) - How did I end up with a kid on the other end of the political spectrum? How did I fail? Steffi, get me a copy of my will... and an eraser.
[... after a long while, right after tumor removed]
(father) - Honey! Bring down a copy of my will... and an eraser! imdb link here
Indeed - Kolmogorov complexity is nice to play with, but can't be calculated.
A useful approximation is to use "compressed size". An ideal, lossless compressor would be readily calculating the kolmogorov complexity. For instance, in the 123456789012345678901234567890 sequence example, any self-respecting compressor such as Zip would create something like
"1234567890 times 3", which is pretty close to the shortest program which generates the sequence.
Indeed, really-good compression is close to AI. To say the same thing in progressively shorter ways, you need to find deeper patterns. Check out this page relating AI to compression: the Hutter prize
Because they draw people to try to reflect their points of view; and when you read the article (say, abortion or evolution or software patents) you can gain a quick overview on almost any significant point of view on the subject, and how they relate to each other. Yes, individual viewpoints may not be perfectly reflected. But you *do* gain an incredibly broad view, which no traditional encyclopedia can deliver.
Wikipedia is much more likely to be useful on a controversial subject where people feel inclined to participate (and correct or refactor partisan views) than in non-controversial subjects that doesn't scratch anybody's itches. You need to cross a certain threshold in order to contribute to an article. Articles that aren't important to you you simply will not edit. Articles that are edited by many may not gain "quality", but will become very broad, and better starting points for further research than those that are only edited by a few not-that-motivated users.
What is this, Digg? This prism group is battling against the govt. demanding copyright for research which they have funded. Whatever you feel about the issue, it has nothing to do with copyright enforcement, but instead copyright ownership. You sir, are probably trolling (unless 'they' refers to 'the government'; then you would be only mistaken). In any case, 'prism' is demanding protection of the old-school scientific publishing business model. A model where
Someone (problably a government grant) pays the researcher to research stuff, and provides the means to do so
The researcher writes a paper describing results, and submits it to a journal
The journal makes sure that the paper's style looks academic enough, and that the topic is inside the journal's range, and looks good enough to be sent to reviewers.
The journal circulates the draft paper among 'referees', which receive no payment whatsoever. This is called peer review.
After a time (typically several months), the journal asks the researcher to revise the paper (which may require another round of peer review), accepts it as-is, or throws it out of the door
Once the paper is published, the journal recoups costs by limiting access to anyone who has not subscribed
The journal spends money to maintain a website where people can subscribe, find and download papers (if subscribed), and do peer review. They also pay editors to do the styling for authors. And contact peer reviewers (referees) and bug them until they perform their reviews. Finally, they decide when something is ready to go to print or not. Then they charge researchers to access the papers.
This made perfect sense in a world of print, where researchers could not possibly produce well-styled papers with typewriters, and manuscripts many times were written by hand, with figures attached inside an envelope. But now every researcher knows how to use MS Word, and most can also use LaTeX, which can easily be made to produce consistent, print-quality papers. Besides, there is no 'printing press' or shipping involved. It only takes a webserver to make your research available to the world. Peer review can be performed on-line, without waiting for referees to check their snail-mail and send the results back again. Suddenly, the role of the great publishing houses has been mostly reduced to that of maintaining a website. Besides the website, the only remaining function of the publisher is to edit. To make the decision of whether something is worthy of peer review or should be kicked out without consideration, and to mediate between reviewers and authors. This role could be taken on by volunteers, universities or much smaller institutions. 'Prism' seems to be worried about the possibility of the US requiring open access to all government-funded research; hence the lobbying.
I wholeheartedly support the open-access movement, and hope that the PLoS initiative gains traction. Unfortunately, my institution still works under the 'publish or perish' model, leading to many poor-quality papers being preferrable (from the career point of view) to few, good ones. And journals without Thomson's ISI blessing are considered pariahs.
On the other hand, many researchers publish their papers on their academic websites. And 'google scholar' does a fine job of finding them there. So, in many cases, publishers are already fighting an uphill battle to stay relevant. Pre-print sites such as Arxiv also provide free access to the meat of many papers, bypassing the publishers.
Slashdot Mirror
But the idea is to be evil while looking innocent. Hard to read != innocent. None of the prize submissions uses easy to find complexity such as weird preprocessor stuff or spaghetti code. They all try to look as professional and straightforward as possible.
For instance, assembly code is right out - writing innocent-looking assembler is way too hard, and people will always suspect it.
Arrays, pointers and functions, no memory protection, dangerous strings. I would like to see the same contest with other 'safer' languages, say Java or Python.
What languages are best suited to underhanded tactics, that is, seemingly innocent but evil?. Notice that underhandedness is very different from plain old abuse -- anybody can write unreadable programs in their favorite language. But, can you make them "clearly read" something different from what is actually written?
Seems like an important question for people who use Open Source because of the difficulty for adding back doors. For many applications, security is at least as important as speed, and you already have The Shootout for that.
I am very interested in your obvious approach to solve shortages in food, potable water, basic shelter, education facilities and farmable land. Especially if it can solve everything at once, instead of focusing on small, well-defined problems.
Negroponte is not superman - but he has a valid point: kids /are/ the future. And to make the future better, you have to invest in education. If you can get kids to learn autonomously and to use computers - well, you're on your way to helping fix all that long list of shortages when the kids grow up.
The kids don't give a damn what the OS is. Hell, if you had never used a computer before, would you know anything about an 'OS' and its religious wars?
True, some people would like to promote OSS with the OLPC. It turns out that OSS is very customizable and has no licensing costs. So it is not such a dumb thing to do; and it got MS to contribute its apps at huge discounts just to not lose on the PR bandwagon. What's not to like? Affordable computer education for kids?
As long as the technology works and there is no recorded RMS greeting during startup... go Linux!
The /why/ is curiosity. Kids have lots of it, but you tend to lose it over time as they get slapped in
the hand and get told by adults to get serious. There's no telling to the number of great engineers (or doctors, or artists, or what-have-you) that we missed out due to stifled curiosity.
If you have a better way to build a mousetrap, build it and see if people will buy it. Trying to tell them they need it before you build one isThe OLPC offers unlimited tinkering and very deep and broad educational (education as in building mental models of things and learning to learn, not as in rote memorization) experience for kids, and can help them learn to read and write and communicate and explore the 'net. It is not "a better mousetrap" - there was no mousetrap before, unless you are referring to the school itself as the mousetrap. And OLPC does not intend to displace schools.
Ok, the business model may not be too sound (but the entry of the ClassMate and 3$ Microsoft software bundles can be seen as partial successes - if the goal is affordable computing to 3rd world kids, things look much brighter than a few years ago). Yes, Negroponte is not a finance magician, and I guess he has learnt the hard way that large corporations do not always place developing nations before shareholder value - that's what PR is for, anyway.
Not to mention that there are simple combinations of moves that will guarantee that the cube will get solved. You can cram those in before an interview and voila! an expert 3D problem solver.
A better test for a programmer would be to say "write me a program that solves rubik's cube"... or if you want to test for 3D prowess, ... graphically.
There will be mafias wherever there are groups or individuals with interests that run contrary to those of the state/society; you can't get rid of crime just by legalizing a few of the currently-illegal interests. Don't want to pay your taxes? Want to get your money back without having to sue the debtor and wait forever? Want something that the owner won't sell?
Not that legalizing marihuana would be negative, I'm just pointing out that crime is here to stay...
Problem is, human troops will usually balk at shooting innocent civilians, or turning on their own country to impose a totalitarian state at the whim of a general. Robot troops would not give a damn, and would obey whoever controls the access codes. Be it president, general, evil genius, or rogue AI.
Yep, we aren't going to run short of apocalyptic science fiction anytime soon...
First, as rdebath argues, you only get 16 bits of CRC on TCP headers.
And furthermore, if you start calculating CRCs off random data, chances (>50%) are you will get a collision (two chunks of data with the same CRC) around the 256th try (this is known as the "birthday paradox" in criptography). Of course, to be really sure to get a collision you will need to try at most 65536 values; but you will reach a very high probability of clash much sooner than intuition may tell you.
See birthday attack for the math.Saying that the ceiling in neural nets has been reached just because no huge breakthroughs have occured lately ignores the fact that our own thinking procecess can very probably be modelled by a sufficiently complex neural net.
We only know how to teach current neural nets a few tricks. But saying that we have reached the end of the road is oversimplification. After all, understanding of thought procecess in the brain is still in its infancy. We don't know how our own "thinking machines" work (which are vastly superior to our current AIs), but we do know that they seem to be based on neurons dynamically signalling each other.
It's not only used by children and politicians; there you have "don't shoot the messenger" (royalty and military), "the ostrich algorithm" (100% effective for technical woes), and "Gandalf Stormcrow" (LOTR version of the above).
So, you're saying that because both Leibniz and Newton independently invented calculus, it must be obvious?
Thousands of college freshmen might disagree with you.
I guess we will be needing lawyers on this one for the forseeable future.
Any system that allows a voter to prove to himself that his vote was correctly cast *after* it has been cast can be perverted into a vote-buying scheme. I could place an stall with an internet connection outside the voting office with a big sign that read "20$ to anybody proving he voted $Party". Or I could be a bit less obvious, and just whisper around the neighborhood that bad things could happen if your slip didn't say what it was supposed to. You get the idea...
So in your qualified opinion, only people that already know unix/linux should use unix/linux.
I think this is great to allow new would-be linux users to install linux with less hassles. Or even experienced users to install linux withouth having to resort to burning a CD or downloading a bootable image into a thumbdrive. Repeat after me: convenience is good. Even if it allows people who shouldn't use it to shoot themselves in the foot. And I bet it can be easily adapted to boot any other distro, perhaps one of the more novice-friendly ones, such as Ubuntu. There's more than one way to use OSS. (Extolling the virtues of OSS on Slashdot: -1 Redundant, here I go :-).
By your line of reasoning, high-level languages should not exist, because people who don't know how to program assembler have no business mucking about with a computer, and dumbing it down is just going to make it easier for them to whine about it.
Germany has about 5 parties above the 2M voters watermark, many of which run local administrations. They can freely organize themselves into coalitions (that is, co-aligned parties can pool their votes together after an election) to achieve majorities when required. Voters have more choices, and there's not that many coalitions a party can enter unless it actually gets voted. Spain also has quite a few major parties. Don't want to vote for Socialist Party (social democrats in europe are like liberals in the US)? -- vote for the United Left party; they have similar goals and are likely to form a coalition if neither can rule by itself. Not that the system is perfect, but you do get slightly more "competition for ideas" and accountability than with a party system where minority parties don't get to play at all. There's lots of ways to organize a parliamentary democracy.
The actual study demonstrated that, for a given population (details on the selection of which were not provided), those that self-described themselves as liberals (whatever that label means) reacted differently to a situation (monitored by a certain measure of neural responses) than those that self-described themselves as conservatives. That part is "good science", though I may find it inconclusive. The bad part is where the study itself (from here) concludes that
Stronger conservatism (versus liberalism) was associated with less neurocognitive sensitivity to response conicts. At the behavioral level, conservatives were also more likely to make errors of commission.This is only true within the experimental population, conservatism and liberalism where asked (instead of "measured"), and the neurocognitive sensitivity can depend on a host of other factors (mostly dependent on the selection of the experimental population), which may better explain the differences between subject's reactions. This part I find wrong.
When you read it, you find this:
a study that found a pattern correlating the way certain people behave and think There may be a pattern that correlates the way people think and their "left vs right" positioning -- but I will refuse to understand that "conservatives just can't help being hard of head" (tempting though it would beAlso, your definition of "liberal" and mine are bound to be quite different; I happen to think it is a gross oversimplification, which fits nicely into a two-party system where many end up thinking there's only two possible stances on each subject (and furthermore, that the one espoused by their "team" is the right one). I think it is obvious that it would be more accurate to quiz people for key issues where "left" and "right" differ, and see where they actually end up when you map the answers to the scale. Or even better, try to correlate with particular stances on particular issues. Asking "conservative or liberal" when neither is clearly defined is, IMHO, flawed.
Your assumption of a .712 on some fictional absolute scale having to have the same stances as another who scored .712 is ridiculous.
I didn't say that. I used it as an example of what's wrong with the simplification. It's like using a "strawberry vs chocolate" metric to find out what you like to eat. Politics should be about taking stances on issues. Political marketing is about convincing you that it's all reducible to "us vs them", and that the only thing you should worry about is cheering your team or bashing the opponents.
It's a matter of how many of your views tend to be liberal or conservative. It's not like if you're a little conservative, you're going to have the same stance on a particular conservative issue as someone else who is a little conservative.
So you say there is a full "conservative" view and a full "liberal" view of the world, and that both of us can agree on what an ideal conservative and an ideal liberal would say on any topic?. That's plain wrong. Hard problems do not have unique "left" and "right" viewpoints.
[...] Political leaning is a part of a human behavior. Everyone leans in one way or the other. Why is it so crazy that those who tend to lean towards one direction tend to think a certain way?
Because I think that oversimplifications are dangerous; and I do not think that the paper supports this particular generalization. I also think that "liberal" and "conservative" are moving targets, and to that extent are more like marketing brands than real products.
[...] On an off-topic note, more than two parties is dangerous because you run the risk of the majority of people not voting for the winner. If you have candidates x, y, & z where y & z share a lot of similar views except one, you can get a situation where 3/5 of the population are split amongst those two. Now, 2/5 of the votes went to x, which 3/5(obviously) of the population did not vote for, but even so, he got more votes than the other two (the other two both getting 3/10 of the vote, whereas (if you can't do math) he got 4/10). That's why more than two political parties is illogical. You can get a winner who can possibly share no views with the majority of the population.
You are describing a "winner takes all" approach. AFAIK, that's very rare. In my country, x&y would have to join forces, or x&z or y&z. Government requires 50% +1 support to be elected. If no agreement is reached, voting is repeated. There's a whole theory on voting out there, and it *is* possible to get multi-party democracies off the ground. In fact, most of the democracies of the world work that way (at least nominally; many are de-facto 2-party systems). Start by checking wipedia.
The bad part about only 2 parties is that you either vote for one, for another, or abstain. You always end up voting "lesser evils", and the system lives on. If there were credible contenders, with which alliances could be formed, you could vote for co-alligned "y" and thus transmit a message to "x" that yes, you are mostly co-aligned, but somehow pissed at them. If there's only "x" and "z", you'd have to either abstain or vote "z". Abstaining and voting blank (marginally better) are too similar, since there are too many ways to interpret what your goal was when doing so. Two-party systems also promote finding where the "middle ground" is and appealing to it (since that is what decides an election if you manage to get everybody nicely layed out in a 1-d spectrum). That gives you reactive, poll-driven politicians which have no real opinions, instead of active politicians who can stand for the ideas you voted them for.
[...]You realize all this study is saying is that those who define themselves as liberals/conservative tend to share a similar way of thinking as others who label themselves as liberals/conserva
Yes, but depending on the college, the prevailing political opinion may be heavily slanted towards one side. That would certainly skew the results, as people who couln't make their minds for themselves would be answering with locally "righteous" ideology, and cases of those who did not cave in would be more extreme (either because they felt strongly about their options, or because they stuck to their choice out of being stubborn). My wild guess is: predominantly liberal college, few conservatives to choose from, most happened to be headstrong.
Repeat the experiment with a different distribution to check for this bias, or quiz people on their political views instead of allowing them to tick a box.
Working in constant touch with researchers, "random populations" in academia are usually whoever happens to be working in or near the lab when the paper's deadline looms near; or students who happen to be in the classroom. Not "random" for any statistically sound definition of the term. For starters, I would have found it more significant if the researcher could specify the distribution of "liberal" vs. "conservative" in the population pool from which the participants where selected. There's places where it is easy to say you're republican, and others where it is easy to say you're liberal.
Also, I do not think that asking people their perceived "liberalness score" actually equates to people standing for the relevant values (and the same could be said of "conservativeness"); as in "I really think this is right" vs. "I say this because it is accepted party dogma".
The linked site is scarce on details - the paper itself is more interesting. First, David Amodio (lead researcher) is not obviously flaming. I'm no expert in neuroscience -- but the data looks good, and he has a track record on usage of scans for similar tasks (most of it race-bias related, but that's another subject).
Here's an interesting part of the experimental design:
I think that there are two ways in which the experiment may be flawed. One is that 43 persons are not enough to extrapolate to the whole US population; and more importantly, no details are given on how they were chosen. If they were chosen among colleagues in an academic setting, where most people (your mileage may vary) are left-wing, you may have problems finding people which self-describe as conservative. These few would be most resistant to changing their viewpoints, I would guess -- since otherwise they may have flipped from exposure to liberal arguments.
Another way in which I think the study may be flawed is by asking people to self-define their position in the political spectrum -- a one-dimensional political spectrum. What guarantees do you have that participants really are "conservative" or "liberal" (whatever that means to you), and have actually thought about the political issues involved in each "choice" (as if there weren't many, many greys)?. A 2-dimensional political positioning would provide more insight. A short questionnaire where participants actually had to think, instead of "choosing their favorite color" would have been even better.
This is assuming that the researcher knows what he's doing, and the
conflict-related ACC activity was indexed by two ERP components. ERPs are scalp-recorded voltage changes reflecting the concerted firing of neurons in response to a psychological event. is actually a good measure of resistance to change or willingness to accomodate it. No details are provided on the exact activity, other than stating that parcicipants were offered the choice of two actions, "Go" and "No-Go".You can find the full article at the author's lab website.
Where the character played by Lukas Haas is originally a rabid republican, until the doctors find that he has a tumor in his head. Once the tumor is removed, he's instantly converted (back) into a left-wing liberal - just like his entire family.
Quotes:
(father) - How did I end up with a kid on the other end of the political spectrum? How did I fail? Steffi, get me a copy of my will... and an eraser.
[... after a long while, right after tumor removed]
(father) - Honey! Bring down a copy of my will... and an eraser!
imdb link here
A useful approximation is to use "compressed size". An ideal, lossless compressor would be readily calculating the kolmogorov complexity. For instance, in the 123456789012345678901234567890 sequence example, any self-respecting compressor such as Zip would create something like "1234567890 times 3", which is pretty close to the shortest program which generates the sequence.
Indeed, really-good compression is close to AI. To say the same thing in progressively shorter ways, you need to find deeper patterns. Check out this page relating AI to compression: the Hutter prize
Because they draw people to try to reflect their points of view; and when you read the article (say, abortion or evolution or software patents) you can gain a quick overview on almost any significant point of view on the subject, and how they relate to each other. Yes, individual viewpoints may not be perfectly reflected. But you *do* gain an incredibly broad view, which no traditional encyclopedia can deliver.
Wikipedia is much more likely to be useful on a controversial subject where people feel inclined to participate (and correct or refactor partisan views) than in non-controversial subjects that doesn't scratch anybody's itches. You need to cross a certain threshold in order to contribute to an article. Articles that aren't important to you you simply will not edit. Articles that are edited by many may not gain "quality", but will become very broad, and better starting points for further research than those that are only edited by a few not-that-motivated users.
The journal spends money to maintain a website where people can subscribe, find and download papers (if subscribed), and do peer review. They also pay editors to do the styling for authors. And contact peer reviewers (referees) and bug them until they perform their reviews. Finally, they decide when something is ready to go to print or not. Then they charge researchers to access the papers.
This made perfect sense in a world of print, where researchers could not possibly produce well-styled papers with typewriters, and manuscripts many times were written by hand, with figures attached inside an envelope. But now every researcher knows how to use MS Word, and most can also use LaTeX, which can easily be made to produce consistent, print-quality papers. Besides, there is no 'printing press' or shipping involved. It only takes a webserver to make your research available to the world. Peer review can be performed on-line, without waiting for referees to check their snail-mail and send the results back again. Suddenly, the role of the great publishing houses has been mostly reduced to that of maintaining a website. Besides the website, the only remaining function of the publisher is to edit. To make the decision of whether something is worthy of peer review or should be kicked out without consideration, and to mediate between reviewers and authors. This role could be taken on by volunteers, universities or much smaller institutions. 'Prism' seems to be worried about the possibility of the US requiring open access to all government-funded research; hence the lobbying.
I wholeheartedly support the open-access movement, and hope that the PLoS initiative gains traction. Unfortunately, my institution still works under the 'publish or perish' model, leading to many poor-quality papers being preferrable (from the career point of view) to few, good ones. And journals without Thomson's ISI blessing are considered pariahs.
On the other hand, many researchers publish their papers on their academic websites. And 'google scholar' does a fine job of finding them there. So, in many cases, publishers are already fighting an uphill battle to stay relevant. Pre-print sites such as Arxiv also provide free access to the meat of many papers, bypassing the publishers.