The people who most directly care about especially quick-to-skim summaries of quality (like impact factor) are people judging the output of professors. If you're not familiar with a sub-field, how do you separate the professor who's published 20 lame papers in questionable venues from the professor who's published 20 high-quality papers in the top journals of his field? You look at some sort of rating for the venues he's published in.
For reading papers, I agree it's not quite as relevant. I still do do a first pass of filtering by using my subjective views of publication quality, though. I'm more likely to give some surprising-sounding claim a thorough evaluation if it was published in a reputable journal than if it was published as a PDF on the internet, or in some obscure conference. You can't read everything, and the well-known conferences and journals in my area provide one level of vetting that I can rely on.
If you want to automatically determine what constitutes a good journal purely from data, the definition is something like: is frequently cited by other good journals. Obviously, there's a circularity there. Various techniques attempt to mitigate it, but none are perfect, and indeed most are rather simplistic and easy to game. It's basically hard to distinguish, purely from citation data, a vibrant community of legitimate research from a vibrant community of crackpots.
In real life, most academics get around the circularity problem by starting with a set of "known good" journals that are determined by consensus in the field rather than algorithms (though this may sometimes be controversial). That lets them take into account more subjective things such as status of a research community (crackpots or not?). For example, as the linked article points out, the Annals of Mathematics is generally accepted as a top-quality venue for mathematics.
If you wanted, you could then construct an Annals-centric view of mathematical impact automatically by seeing how frequently other journals are cited by papers in Annals. This is what happens informally as journals gain and lose reputation: a promising new venue often first comes to a community's attention because its articles begin to be cited in "known good" journals.
But just taking all journals with no starting point, and attempting to extract from the citation graph which ones are "good" purely from the links, is doomed to failure, because there just isn't enough information in there to make the distinctions people want to make.
Many long-tail proponents argue that the two meanings you discuss are related: when distribution/stocking costs for obscure works go to zero, it can be profitable to stock/sell them, and therefore they will actually be stocked/sold by online retailers. Since this makes the works much more readily available than previously, their sales will (collectively) go up.
The hypothesis is that the reason the top N works account for such a large market share traditionally is at least in part the practical problems with stocking more than a certain amount of stuff at your local retailer. Therefore, the hypothesis goes, obscure works will collectively account for a larger percentage of sales once the shelf-space constraint goes away and people can buy what they would "naturally" prefer.
The present article seems to be arguing that, even though obscure works are now easier to get, they're actually still not really being bought much. The article seems to be arguing that the constraining factor is now a sort of virtual shelf space of mind-share/visibility/recommendations, so the obscure works don't get noticed even though they are now easy to find and buy if you knew about them.
In the climate of the past few years, it's been mainly Democrats objecting to Bush-administration War-on-Terror measures. But it's also of course a libertarian position.
There are games that plausibly serve some sort of editorializing function, and then there are games that just reference recent events, usually as a gimmick. Many of the recent Bush-shoe-throwing games are of the second sort---there is no real editorial commentary going on, it's just a generic arcade game that's been skinned with Bush and shoes. There were similarly content-free games that came out after 9/11, mostly based around revenge fantasies where you got to punch bin Laden or something.
There are some good examples of games that actually use the gameplay to make some sort of editorial point, though. From a right-wing perspective, in Al Quaidamon, you can treat a terrorist prisoner well or poorly, and a meter shows his current status. The political point is made in the balance: unless you coddle him continuously, you fall below the levels market as Geneva Convention standards (which are, incidentally, depicted as being above average U.S. living standards). From a more left-wing perspective, Airport Security satirizes the post-9/11 airport security measures through its gameplay, by depicting the changing standards of what's banned this week as absurd and impossible to follow.
In the academic field of game studies (analogous to film studies, though much smaller), the idea of games as rhetoric/etc. has been discussed for several years. Probably the most prominent academic who also makes games in that vein is Ian Bogost, who explicitly describes a lot of what he does as making "playable editorial cartoons". The New York Times for a while was actually publishing them on its online editorial page, strengthening the analogy (until a change of editor). He also happens to have a book on the somewhat broader subject of games as a means of commentary/expression/rhetoric, Persuasive Games: The Expressive Power of Videogames (MIT Press, 2007).
There is also an index here of editorial-style "newsgames", i.e. games about recent news events released in a timely manner that make some editorial commentary about the event.
Script suggestion: Have someone saying "I'm Linux", yelling starts off-camera, camera pans over sort of haphazardly, and Stallman launches into a rant about how it's GNU/Linux.
You're seriously arguing that it would be an intelligent move to accept a manual-labor, outdoor construction job for $10/hr, when you could get paid $8/hr to work from your bedroom?
Why do you think LBJ, for example, kept Hoover and his cronies (of which Felt was one) in office? To keep them from releasing dirt on him, which they duly didn't. The Hoover/Felt mafia dished the dirt as soon as someone stopped paying the protection money, which was 1972.
it's not so much that there are blind spots
on
Trick or Treatment
·
· Score: 2, Informative
It's that the entire premise this book's authors are coming from---that standard medicine is about evidence-based medicine---is not really universally accepted in standard medicine. Its acceptance is growing, but EBM as an explicit aim was only introduced in the early 1990s, and was initially seen as basically a crusade by a bunch of ivory-tower lab scientists who didn't understand the subjective complexity of real-world clinical practice. It's only from the late 1990s or so seen increasing acceptance in affecting clinical practice.
So to some extent, saying "homeopathy is wrong because it doesn't follow EBM principles" is a bit off target, because it's not the primary thing that distinguishes standard from alternative medicine.
I should add that it is improving
on
Trick or Treatment
·
· Score: 2, Informative
Especially in areas where there's some specific push to use evidence-based medicine, its adoption is increasing and leading slowly to changes in clinical practice, as long-established assumptions have turned out not to be supportable by evidence.
One of the more notable examples is the significant decrease in use of antibiotics for many bacterial maladies, which has been driven by an initiative to experimentally validate allegedly positive uses of antibiotics, and stop prescribing them if evidence of positive effect can't be found.
It used to be assumed that, because broad-spectrum antibiotics kill bacteria, they are therefore useful to prescribe for maladies caused by bacteria. However in many cases they turn out to have little effect at all; for example, controlled studies of antibiotic prescription for ear infections have generally shown no improvement in recovery speed or likelihood with antibiotics as compared to without. Therefore the previous, non-evidence-based standard medical practice ("you should prescribe antibiotics for ear infections") has turned out not to be experimentally supportable.
A lot of standard medicine doesn't really pass the test of evidence-based medicine either, in the sense that specific advocated treatments have been validated experimentally when applied to specific, observable conditions. That's one reason EBM is still relatively controversial: many standard surgical and medical practices are based on rational inferences from facts we're pretty sure of, but have never themselves been validated.
To take a really simple example, look at how dermatologists treat moles. There isn't very good experimental data on mole prognosis. An EBM approach would say something like: given specific observed features of this mole, data tells us it has an x% chance of turning into a melanoma within Y years. You would probably need computer models to aggregate the various features that could contribute to or against it being at risk. Dermatologists don't generally have this information at hand (if it exists at all), but instead make more subjective judgment calls, based on some high-level knowledge of risk factors (which may or may not have ever been validated experimentally themselves).
You don't have to think environmental impact assessments are a bad thing to agree that they're a major reason there are no battery factories being built in the US. Battery factories are very dirty, at least using current production methods, and possibly inherently at least questionable (there are a lot of heavy metals and whatnot going into them).
The biggest problem with unions from my perspective is that in many fields they curtail my ability to earn good wages, rather than help it, by prohibiting people with in-demand skills from being offered wages or salaries outside of a rigidly specified pay structure.
For example, in a free market I could earn much more money as a computer-science schoolteacher than the union will allow me to make---because most teachers' union contracts require that all subjects be paid the same. Even though science, math, and CS teachers are much more in demand than social-studies teachers, the district cannot offer higher salaries to attract them. This directly harms science/math/CS teachers, who to add insult to injury, are still required to join the union (though I know several who have forced the union to take them to court rather than join willingly).
If the unions don't make outright efforts to harm workers, they might encounter less resistance.
Corporate taxes are certainly not a cost "just like any other cost". If they were revenue taxes, that would be true, but they're taxes only on profits, which leads to quite different effects from most costs. If Southwest makes no profit this year, they pay zero in corporate tax, but you aren't going to see their fuel costs going down analogously.
In effect a corporate tax reduces any positive profits proportionally to how many positive profits there are. The main negative effect this has on business is reducing incentive of entrepreneurs to go into business; certainly past some level, where it reduced non-taxed profit almost to zero, it would stifle innovation entirely. But if a market segment is very profitable, then a corporate tax really has no effect except reducing profits---because the market segment will still be very profitable even after the tax.
Of course, one problem with analyzing corporate taxes using neoclassical price theory, as you seem to be attempting, is that in neoclassical price theory profits don't exist: in an efficient competitive market, the prices of goods and services approach their marginal cost of production. In that view, corporate taxes are basically a tax on market inefficiencies.
Suicidal people aren't 100% irrational zombies or something. They seize on things and overemphasize them, downplay contrary evidence, etc., but they do still have thought processes that take into account the external world.
One of the (many) ways of trying to convince people who are in particular suicidal because of a desire to "get back" at someone is that suicide is not a particularly effective way of getting back at people. Providing a very concrete way in which it arguably actually is a good way of getting back at someone is not very helpful from that perspective.
Judging from this post, it looks like the changes involved support for mixed Windows/Linux calling conventions on x86-64 (i.e. specifying on a per-function basis whether to use the Windows or Linux calling convention).
Google's services all also, of course, "introspect on incoming data in order to automatically classify it and extract structure from it using machine learning" -- for the purpose of serving up contextual ads.
The main point of the article as I read it is that it makes more sense to treat netbooks as powerful PDAs rather than as weak laptops. People don't really expect their smartphones, for example, to run Windows apps, but instead expect them to have some useful functionality built in that is not too hard to figure out. Seen that way, a netbook with a web-browser, decent office package, and some camera-syncing software built in ought to be sufficient for many people, especially those who do much of their "real" computing on the web (use gmail, manage their pictures via flickr, etc.).
The target market for netbooks is generally "normal people", who are more or less by definition not editing Word documents with mathematical formulas in them.
Mining is hugely polluting, both to landscapes and to water, especially the sorts of mining that batteries need (heavy metals). If we had to ramp up production to populate the entire planet with many-battery-containing vehicles, that would be a significant effect.
It would also require a very well run and credible disposal/recycling program at the end of life, i.e. not just ship the dead batteries off to China.
For desktops that might be acceptable, but shoving a 15,000-rpm drive in a laptop is going to kill the battery life compared to an SSD, not to mention increase the noise output and heat-dissipation issues.
Slashdot Mirror
The people who most directly care about especially quick-to-skim summaries of quality (like impact factor) are people judging the output of professors. If you're not familiar with a sub-field, how do you separate the professor who's published 20 lame papers in questionable venues from the professor who's published 20 high-quality papers in the top journals of his field? You look at some sort of rating for the venues he's published in.
For reading papers, I agree it's not quite as relevant. I still do do a first pass of filtering by using my subjective views of publication quality, though. I'm more likely to give some surprising-sounding claim a thorough evaluation if it was published in a reputable journal than if it was published as a PDF on the internet, or in some obscure conference. You can't read everything, and the well-known conferences and journals in my area provide one level of vetting that I can rely on.
If you want to automatically determine what constitutes a good journal purely from data, the definition is something like: is frequently cited by other good journals. Obviously, there's a circularity there. Various techniques attempt to mitigate it, but none are perfect, and indeed most are rather simplistic and easy to game. It's basically hard to distinguish, purely from citation data, a vibrant community of legitimate research from a vibrant community of crackpots.
In real life, most academics get around the circularity problem by starting with a set of "known good" journals that are determined by consensus in the field rather than algorithms (though this may sometimes be controversial). That lets them take into account more subjective things such as status of a research community (crackpots or not?). For example, as the linked article points out, the Annals of Mathematics is generally accepted as a top-quality venue for mathematics.
If you wanted, you could then construct an Annals-centric view of mathematical impact automatically by seeing how frequently other journals are cited by papers in Annals. This is what happens informally as journals gain and lose reputation: a promising new venue often first comes to a community's attention because its articles begin to be cited in "known good" journals.
But just taking all journals with no starting point, and attempting to extract from the citation graph which ones are "good" purely from the links, is doomed to failure, because there just isn't enough information in there to make the distinctions people want to make.
Many long-tail proponents argue that the two meanings you discuss are related: when distribution/stocking costs for obscure works go to zero, it can be profitable to stock/sell them, and therefore they will actually be stocked/sold by online retailers. Since this makes the works much more readily available than previously, their sales will (collectively) go up.
The hypothesis is that the reason the top N works account for such a large market share traditionally is at least in part the practical problems with stocking more than a certain amount of stuff at your local retailer. Therefore, the hypothesis goes, obscure works will collectively account for a larger percentage of sales once the shelf-space constraint goes away and people can buy what they would "naturally" prefer.
The present article seems to be arguing that, even though obscure works are now easier to get, they're actually still not really being bought much. The article seems to be arguing that the constraining factor is now a sort of virtual shelf space of mind-share/visibility/recommendations, so the obscure works don't get noticed even though they are now easy to find and buy if you knew about them.
In the climate of the past few years, it's been mainly Democrats objecting to Bush-administration War-on-Terror measures. But it's also of course a libertarian position.
There are games that plausibly serve some sort of editorializing function, and then there are games that just reference recent events, usually as a gimmick. Many of the recent Bush-shoe-throwing games are of the second sort---there is no real editorial commentary going on, it's just a generic arcade game that's been skinned with Bush and shoes. There were similarly content-free games that came out after 9/11, mostly based around revenge fantasies where you got to punch bin Laden or something.
There are some good examples of games that actually use the gameplay to make some sort of editorial point, though. From a right-wing perspective, in Al Quaidamon, you can treat a terrorist prisoner well or poorly, and a meter shows his current status. The political point is made in the balance: unless you coddle him continuously, you fall below the levels market as Geneva Convention standards (which are, incidentally, depicted as being above average U.S. living standards). From a more left-wing perspective, Airport Security satirizes the post-9/11 airport security measures through its gameplay, by depicting the changing standards of what's banned this week as absurd and impossible to follow.
(I got both of those examples from this list.)
In the academic field of game studies (analogous to film studies, though much smaller), the idea of games as rhetoric/etc. has been discussed for several years. Probably the most prominent academic who also makes games in that vein is Ian Bogost, who explicitly describes a lot of what he does as making "playable editorial cartoons". The New York Times for a while was actually publishing them on its online editorial page, strengthening the analogy (until a change of editor). He also happens to have a book on the somewhat broader subject of games as a means of commentary/expression/rhetoric, Persuasive Games: The Expressive Power of Videogames (MIT Press, 2007).
There is also an index here of editorial-style "newsgames", i.e. games about recent news events released in a timely manner that make some editorial commentary about the event.
Stallman would never say "I'm Linux". ;-)
Script suggestion: Have someone saying "I'm Linux", yelling starts off-camera, camera pans over sort of haphazardly, and Stallman launches into a rant about how it's GNU/Linux.
than the Compaq Evo
You're seriously arguing that it would be an intelligent move to accept a manual-labor, outdoor construction job for $10/hr, when you could get paid $8/hr to work from your bedroom?
Why do you think LBJ, for example, kept Hoover and his cronies (of which Felt was one) in office? To keep them from releasing dirt on him, which they duly didn't. The Hoover/Felt mafia dished the dirt as soon as someone stopped paying the protection money, which was 1972.
It's that the entire premise this book's authors are coming from---that standard medicine is about evidence-based medicine---is not really universally accepted in standard medicine. Its acceptance is growing, but EBM as an explicit aim was only introduced in the early 1990s, and was initially seen as basically a crusade by a bunch of ivory-tower lab scientists who didn't understand the subjective complexity of real-world clinical practice. It's only from the late 1990s or so seen increasing acceptance in affecting clinical practice.
So to some extent, saying "homeopathy is wrong because it doesn't follow EBM principles" is a bit off target, because it's not the primary thing that distinguishes standard from alternative medicine.
Especially in areas where there's some specific push to use evidence-based medicine, its adoption is increasing and leading slowly to changes in clinical practice, as long-established assumptions have turned out not to be supportable by evidence.
One of the more notable examples is the significant decrease in use of antibiotics for many bacterial maladies, which has been driven by an initiative to experimentally validate allegedly positive uses of antibiotics, and stop prescribing them if evidence of positive effect can't be found.
It used to be assumed that, because broad-spectrum antibiotics kill bacteria, they are therefore useful to prescribe for maladies caused by bacteria. However in many cases they turn out to have little effect at all; for example, controlled studies of antibiotic prescription for ear infections have generally shown no improvement in recovery speed or likelihood with antibiotics as compared to without. Therefore the previous, non-evidence-based standard medical practice ("you should prescribe antibiotics for ear infections") has turned out not to be experimentally supportable.
A lot of standard medicine doesn't really pass the test of evidence-based medicine either, in the sense that specific advocated treatments have been validated experimentally when applied to specific, observable conditions. That's one reason EBM is still relatively controversial: many standard surgical and medical practices are based on rational inferences from facts we're pretty sure of, but have never themselves been validated.
To take a really simple example, look at how dermatologists treat moles. There isn't very good experimental data on mole prognosis. An EBM approach would say something like: given specific observed features of this mole, data tells us it has an x% chance of turning into a melanoma within Y years. You would probably need computer models to aggregate the various features that could contribute to or against it being at risk. Dermatologists don't generally have this information at hand (if it exists at all), but instead make more subjective judgment calls, based on some high-level knowledge of risk factors (which may or may not have ever been validated experimentally themselves).
[*mumbles under his breath and waves cane threateningly*]
You don't have to think environmental impact assessments are a bad thing to agree that they're a major reason there are no battery factories being built in the US. Battery factories are very dirty, at least using current production methods, and possibly inherently at least questionable (there are a lot of heavy metals and whatnot going into them).
The traffic levels aren't even close.
The fear is that they won't.
The biggest problem with unions from my perspective is that in many fields they curtail my ability to earn good wages, rather than help it, by prohibiting people with in-demand skills from being offered wages or salaries outside of a rigidly specified pay structure.
For example, in a free market I could earn much more money as a computer-science schoolteacher than the union will allow me to make---because most teachers' union contracts require that all subjects be paid the same. Even though science, math, and CS teachers are much more in demand than social-studies teachers, the district cannot offer higher salaries to attract them. This directly harms science/math/CS teachers, who to add insult to injury, are still required to join the union (though I know several who have forced the union to take them to court rather than join willingly).
If the unions don't make outright efforts to harm workers, they might encounter less resistance.
Corporate taxes are certainly not a cost "just like any other cost". If they were revenue taxes, that would be true, but they're taxes only on profits, which leads to quite different effects from most costs. If Southwest makes no profit this year, they pay zero in corporate tax, but you aren't going to see their fuel costs going down analogously.
In effect a corporate tax reduces any positive profits proportionally to how many positive profits there are. The main negative effect this has on business is reducing incentive of entrepreneurs to go into business; certainly past some level, where it reduced non-taxed profit almost to zero, it would stifle innovation entirely. But if a market segment is very profitable, then a corporate tax really has no effect except reducing profits---because the market segment will still be very profitable even after the tax.
Of course, one problem with analyzing corporate taxes using neoclassical price theory, as you seem to be attempting, is that in neoclassical price theory profits don't exist: in an efficient competitive market, the prices of goods and services approach their marginal cost of production. In that view, corporate taxes are basically a tax on market inefficiencies.
Suicidal people aren't 100% irrational zombies or something. They seize on things and overemphasize them, downplay contrary evidence, etc., but they do still have thought processes that take into account the external world.
One of the (many) ways of trying to convince people who are in particular suicidal because of a desire to "get back" at someone is that suicide is not a particularly effective way of getting back at people. Providing a very concrete way in which it arguably actually is a good way of getting back at someone is not very helpful from that perspective.
Judging from this post, it looks like the changes involved support for mixed Windows/Linux calling conventions on x86-64 (i.e. specifying on a per-function basis whether to use the Windows or Linux calling convention).
Google's services all also, of course, "introspect on incoming data in order to automatically classify it and extract structure from it using machine learning" -- for the purpose of serving up contextual ads.
The main point of the article as I read it is that it makes more sense to treat netbooks as powerful PDAs rather than as weak laptops. People don't really expect their smartphones, for example, to run Windows apps, but instead expect them to have some useful functionality built in that is not too hard to figure out. Seen that way, a netbook with a web-browser, decent office package, and some camera-syncing software built in ought to be sufficient for many people, especially those who do much of their "real" computing on the web (use gmail, manage their pictures via flickr, etc.).
The target market for netbooks is generally "normal people", who are more or less by definition not editing Word documents with mathematical formulas in them.
Mining is hugely polluting, both to landscapes and to water, especially the sorts of mining that batteries need (heavy metals). If we had to ramp up production to populate the entire planet with many-battery-containing vehicles, that would be a significant effect.
It would also require a very well run and credible disposal/recycling program at the end of life, i.e. not just ship the dead batteries off to China.
For desktops that might be acceptable, but shoving a 15,000-rpm drive in a laptop is going to kill the battery life compared to an SSD, not to mention increase the noise output and heat-dissipation issues.