I wouldn't quite say shot down. It is the subject of legitimate debate. The fact is that garbage collection is competitive with stack allocation, that manual allocation and free are error-prone and inefficient.
Boehm is an advocate of in-place automatic collection, as opposed to stack allocation.
If you care to point to one of the references that "shoots down" Appel, I'd be prepared to comment on it. Otherwise its just hearsay, isn't it?
AI has been oversold time and time again./. participates in publicizing the extravagent predictions of AI-ers.
There has been tremendous progress in building software and hardware systems to do things that previously only humans could do well. Chess is an example; so is spam detection; so are various forms of pattern recognition.
Where AI efforts have been singularly unsuccessful at is in replacing humans entirely for complex tasks in an unpredictable envirnoment. Also to do anything resembling "understanding." Yet proposals to do just this draw the biggest military grants.
Expect the fearless predictions to continue. The grandiose proposals will continue to draw funds, but those funds will be wasted pending the occurrence of some monumental flash of insight.
It is true that the Chinese censorship has many holes. When I was in Shanghai last April, I was able to tunnel to a western server and be censorship-free. On the other hand, other than a few annoyances, like having google.ca redirected to google.cn, and Groklaw (!) blocked, I couldn't find any inflammatory stuff that I couldn't reach directly. Of course, I don't read Chinese, so non-English pages might be their principal target.
But one should not take the existence of loopholes to mitigate the perniciousness of censorship. The educated/motivated people who know
the workarounds are not the issue. The purpose of this censorship is
to prevent the unwashed masses from being exposed to "subversive" ideas.
At that I fear it may be effective.
Bottom line: don't be complacent just because it doesn't stop *you*
from reading your favourite web pages.
The 2004 world final was a bit of a marathon. I'm not sure I'd say that any old solution would do, but the contest was overburdened with a large number of tedious case-analysis and geometry problems. It lacked "algorithms" problems.
I think there's general consensus among the coaches that the 2005 problem set was more appealing. It had brain teasers and algorithms. The top 4 teams were struggling with 3 different problems in the last hour. The winner, Shangai, happened to be the ones who figured out the problem they tackled, not the fastest typers.
Let's hope that 2004 was an anomaly and future contests are more like 2005.
By "these competitions" do you mean TopCoder? IOI and ACM are both 5 hours long. While typing speed may be a bit of an asset in these, it is a minor consideration.
I wish Slashdot would not parrot press releases that contain no information.
Grammatical inference is a hard problem. A couple of researchers, who appear to have legitimate credentials, have published a paper (that's what they do for a living) and acquired a patent (see any number of threads for how much novelty that may or may not require).
I would be much more impressed if there were at least some whiff of a scientific claim, like "after processing 1GB of English text for 10000 hours, the prototype implementation generated text that 51.7% of observers could not distinguish from a transcript of English spoken by a 6-year-old."
It is impossible to determine what, if any, contribution to natural language processing is included in this paper. Therefore the press release is not news.
If somebody wants to *read* the paper and tell us what it is about, I'll be all ears.
The Tragically Hip are correct. New Orleans is sinking, and will continue to sink.
The land is a flood plain. It depends on annual Mississippi flooding to deposit silt and moisture to maintain the land mass. The river levees cut off this replenishment and the land sinks.
The problem will only get worse, and there's no obvious solution.
Incidentally, I've been working with Gordon Cormack to try and figure out what the heck went wrong with his first set of dspam tests. So far, we've made progress and ran a successful test with an overall accuracy of 99.23% (not bad for a simulation).
First, I would like to thank Jonathan for his recent helpful correspondence in configuring DSPAM for the TREC Spam Filter Evaluation Tool Kit. When finalized, this configuration will replace the one currently available (along with seven others).
However, I take exception to the statement above, implying that there is something wrong with the tests that Lynam and I previously published. I stand by those results. Since that report was made public, I have become aware of two others that achieve much the same results: Holden's Spam Filtering II and Sergeant's CRM114 and DSPAM (Virus Bulletin, no longer freely available).
Lynam and I said that DSPAM 2.8.3, in its default configuration, achieved 98.15% accuracy on the same corpus to which Zdziarski refers above. The report also argued that accuracy was a very poor measure of filter performance and that a false positive rate such as the 1.28% demonstrated by DSPAM would likely be unacceptable to an email user.
In recent correspondence, Zdziarski suggested three configurations of DSPAM (available here on the web) that achieved the following results:
More detailed results and comparisons will be made available when our current study is complete.
Don't take my word (or Jonathan's) for anything; run this filter and others on your own email. But please take great care in constructing your gold standard.
I dunno. My dad (78 yrs old) converted to Linux. He uses Linux stuff: Mozilla, OpenOffice, The Gimp, K3B, and so on.
I installed wine and all the apps he was used to -- MS Office, Photoshop, IE (actually he dumped IE for Mozilla on Windows a couple of years ago.) -- but he didn't really use them. He found it easy to switch. In the case of K3B he said it was "much easier to use than that Roxy-whatever thing on Windows." His slide scanner worked perfectly without the installation of any extra software, and mult-vendor multimedia/DVD "just works" unlike in windows.
So I guess that having the Windows apps there provides comfort. But in this particular case study, they were placebos.
Is the article's characterization "grounded" based on anything more substantial than the near-tautology "Until we're ready, we won't fly" from the cited article?
The same could've been said about the fuel sensor malfunction.
Your use of "philosophical" as a pejorative adds little to the discussion. For what it is worth, here's OED's definition:
(= metaphysical philosophy.) That department of knowledge or study which
deals with ultimate reality, or with the most general causes and principles of
things. (Now the most usual sense.)
You go on to say that evolution has no backing other than "it is observed."
One must distinguish between prior and posterior observations. Prior observations are those used to formulate the theory. One can evaluate the elegance of the theory by how well it covers prior observations, but one cannot evaluate its validity.
On the other hand, posterior observations - ones predicted by the theory but not previously observed, either contradict or lend support to the theory. There are any number of posterior observations that support the theory of evolution.
Posterior observations may relate to prior events, such as the fossil record, or to future experiments such as those conducted on fruit flies . They all provide physical evidence that supports the theory. So much evidence that the theory may be considered valid beyond reasonable doubt.
On the other hand, there are not, and I daresay never will be, any posterior observations that support Intelligent Design.
That's how science - and indeed philosphy - works. In more common terms, Put up or shut up.
"but that definition would also fit intelligent design."
I don't agree; however, who cares?
You're the one that introduced the word play in an effort to equate evolution with speculation. If you want to include speculation under the definition of "theory," feel free, but your interpretation *still* says nothing about evolution.
Scientists ever strive for better theories; that is, simpler models built from well-understood concepts and tools that better predict natural phenomena.
Intelligent Design, if it is a theory, is a lousy one. It relies on faith rather than understanding, and it predicts nothing. At least nothing in the natural world.
The terms are pretty well self-explanatory, though they are inessential elements of evolution. For evolution to occur, all we have to have is a reproductive advantage for organisms with some feature. Survival is, of course, a good step to reproductive advantage. But so is being more attractive to or attracted to the opposite sex, shorter gestation, multiple births, and any number of other factors.
If an organism acquires, by chance, a feature that gives it and its offspring a reproductive advantage that results in even a tiny increase in net fertility rate, its offspring will prosper relative to those that don't have it. Over many generations the feature will become common.
The article addresses the issue of when a particular set of features differentiates organisms sufficiently that species separation occurs. As has been pointed out previously, what is at issue is the details of the mechanism, not its existence.
"A theory is a theory" is gobbledegook. It is meant to imply that scientific theory is mere conjecture, which it is not.
A theory, according to the OED, is:
A scheme or system of ideas or statements held as an explanation or account of a group of facts or phenomena; a hypothesis that has been confirmed or established by observation or experiment, and is propounded or accepted as accounting for the known facts; a statement of what are held to be the general laws, principles, or causes of something known or observed.
You don't understand the concept of "fair and balanced." It means that for every person who expresses an accepted and scientifically justifiable opinion, you give equal or greater weight to selected whackos who disagree.
Then, once it becomes accepted that there is "no consensus" you split the difference, and find some even more extreme whackos to skew the "middle ground" even further. Eventually those with well-considered opinion are completely marginalized.
An honest scientist cannot win in this environment, because he or she is not willing to take ever-more-extreme positions to maintain "balance".
I've done a lot of mainframe development and a lot of Unix/Linux development; scarcely any Windows.
The main difference I see between mainframe development and *ix development culture is respect. With the mainframe you have to book time days in advance and work in the wee hours to make any changes. And you make damned sure that, when you're done, things work as they should.
With *ix development, things are laissez-faire. You send out a message a few hours/days/minutes in advance of some monumental change. Then you blame the users when they can't sign on to their system in the morning. Quote some recently-adopted standards if they argue.
Of course, I'm speaking of the early days of *ix. These systems are more and more critical, and the admins are trying to learn respect. But they're playing catch-up. There's nothing like the fear of taking down a $500/minute system to make you careful.
Windows development follows a similar pattern. The whole culture is so "personal computer" based that the concept of a year's continuous uptime is foreign.
The book describes magnetic core which was pretty well obsolete in 1971. It describes CPU, disk, and tape that have become smaller and faster, but that's about it. The monumental developments in programming languages - object-oriented, functional programming, concurrent programming - were in the past. Ditto for file systems, data structures, and operating systems.
If you learned computer architecture back then, you wouldn't have much difficulty with today. Not like the man-frozen-in-the-glacier movie scenario at all.
The development that wasn't even a gleam in the eye of computer scientists of the day is networking.
When you inflate a balloon you supply energy which is used to displace air in the atmosphere. There's no free lunch.
My point about EV was that, whatever mechanism you use, you'll need at least that much energy.
Suppose you rode your helium balloon to the moon and then cut your basket loose. By the time you fell to earth you'd have reached escape velocity. If you really could get to the moon with less energy you'd have invented a perpetual motion machine.
Slashdot Mirror
I wouldn't quite say shot down. It is the subject of legitimate debate. The fact is that garbage collection is competitive with stack allocation, that manual allocation and free are error-prone and inefficient.
Boehm is an advocate of in-place automatic collection, as opposed to stack allocation.
If you care to point to one of the references that "shoots down" Appel, I'd be prepared to comment on it. Otherwise its just hearsay, isn't it?
Rather than promoting your prejudice, why not RTFA, or this 19-year-old paper.
I'll await your learned rebuttal.
AI has been oversold time and time again. /. participates in publicizing the extravagent predictions of AI-ers.
There has been tremendous progress in building software and hardware systems to do things that previously only humans could do well. Chess is an example; so is spam detection; so are various forms of pattern recognition.
Where AI efforts have been singularly unsuccessful at is in replacing humans entirely for complex tasks in an unpredictable envirnoment. Also to do anything resembling "understanding." Yet proposals to do just this draw the biggest military grants.
Expect the fearless predictions to continue. The grandiose proposals will continue to draw funds, but those funds will be wasted pending the occurrence of some monumental flash of insight.
But one should not take the existence of loopholes to mitigate the perniciousness of censorship. The educated/motivated people who know the workarounds are not the issue. The purpose of this censorship is to prevent the unwashed masses from being exposed to "subversive" ideas. At that I fear it may be effective.
Bottom line: don't be complacent just because it doesn't stop *you* from reading your favourite web pages.
I would never exempt them, especially if if "charity" includes for-profit peddlars that give 1% of the take to charity for the use of its name.
However, I have no objection to allowing others to exempt charities, if it will get the bill passed.
If you RTFA, it is proposed opposition amendments that would destroy the no-call list. These amendments have zero chance of becoming law.
The proposed government amendment - to allow subscribers to exempt charities - sounds fine to me.
The 2004 world final was a bit of a marathon. I'm not sure I'd say that any old solution would do, but the contest was overburdened with a large number of tedious case-analysis and geometry problems. It lacked "algorithms" problems.
I think there's general consensus among the coaches that the 2005 problem set was more appealing. It had brain teasers and algorithms. The top 4 teams were struggling with 3 different problems in the last hour. The winner, Shangai, happened to be the ones who figured out the problem they tackled, not the fastest typers.
Let's hope that 2004 was an anomaly and future contests are more like 2005.
By "these competitions" do you mean TopCoder? IOI and ACM are both 5 hours long. While typing speed may be a bit of an asset in these, it is a minor consideration.
I wish Slashdot would not parrot press releases that contain no information.
Grammatical inference is a hard problem. A couple of researchers, who appear to have legitimate credentials, have published a paper (that's what they do for a living) and acquired a patent (see any number of threads for how much novelty that may or may not require).
I would be much more impressed if there were at least some whiff of a scientific claim, like "after processing 1GB of English text for 10000 hours, the prototype implementation generated text that 51.7% of observers could not distinguish from a transcript of English spoken by a 6-year-old."
It is impossible to determine what, if any, contribution to natural language processing is included in this paper. Therefore the press release is not news.
If somebody wants to *read* the paper and tell us what it is about, I'll be all ears.
http://www.azlyrics.com/lyrics/tragicallyhip/newor leansissinking.html
I think he means something like The TREC Spam Filter Evaluation Toolkit.
The Tragically Hip are correct. New Orleans is sinking, and will continue to sink.
The land is a flood plain. It depends on annual Mississippi flooding to deposit silt and moisture to maintain the land mass. The river levees cut off this replenishment and the land sinks.
The problem will only get worse, and there's no obvious solution.
Here's the learning curve and the ROC curve.
Lynam and I said that DSPAM 2.8.3, in its default configuration, achieved 98.15% accuracy on the same corpus to which Zdziarski refers above. The report also argued that accuracy was a very poor measure of filter performance and that a false positive rate such as the 1.28% demonstrated by DSPAM would likely be unacceptable to an email user.
In recent correspondence, Zdziarski suggested three configurations of DSPAM (available here on the web) that achieved the following results:
dspam(tum) fpr 1.81% fnr 0.80% accuracy 99.20%
dspam(toe) fpr 1.94% fnr 0.59% accuracy 99.16%
dspam(teft) frp 1.85% fnr 0.53% accuracy 99.32%
More detailed results and comparisons will be made available when our current study is complete. Don't take my word (or Jonathan's) for anything; run this filter and others on your own email. But please take great care in constructing your gold standard.
Gordon Cormack
Epson 1200U flatbed with B81317 TPU attachment.
I dunno. My dad (78 yrs old) converted to Linux. He uses Linux stuff: Mozilla, OpenOffice, The Gimp, K3B, and so on.
I installed wine and all the apps he was used to -- MS Office, Photoshop, IE (actually he dumped IE for Mozilla on Windows a couple of years ago.) -- but he didn't really use them. He found it easy to switch. In the case of K3B he said it was "much easier to use than that Roxy-whatever thing on Windows." His slide scanner worked perfectly without the installation of any extra software, and mult-vendor multimedia/DVD "just works" unlike in windows.
So I guess that having the Windows apps there provides comfort. But in this particular case study, they were placebos.
Is the article's characterization "grounded" based on anything more substantial than the near-tautology "Until we're ready, we won't fly" from the cited article?
The same could've been said about the fuel sensor malfunction.
One must distinguish between prior and posterior observations. Prior observations are those used to formulate the theory. One can evaluate the elegance of the theory by how well it covers prior observations, but one cannot evaluate its validity.
On the other hand, posterior observations - ones predicted by the theory but not previously observed, either contradict or lend support to the theory. There are any number of posterior observations that support the theory of evolution.
Posterior observations may relate to prior events, such as the fossil record, or to future experiments such as those conducted on fruit flies . They all provide physical evidence that supports the theory. So much evidence that the theory may be considered valid beyond reasonable doubt.
On the other hand, there are not, and I daresay never will be, any posterior observations that support Intelligent Design.
That's how science - and indeed philosphy - works. In more common terms, Put up or shut up.
"but that definition would also fit intelligent design."
I don't agree; however, who cares?
You're the one that introduced the word play in an effort to equate evolution with speculation. If you want to include speculation under the definition of "theory," feel free, but your interpretation *still* says nothing about evolution.
Scientists ever strive for better theories; that is, simpler models built from well-understood concepts and tools that better predict natural phenomena.
Intelligent Design, if it is a theory, is a lousy one. It relies on faith rather than understanding, and it predicts nothing. At least nothing in the natural world.
The terms are pretty well self-explanatory, though they are inessential elements of evolution. For evolution to occur, all we have to have is a reproductive advantage for organisms with some feature. Survival is, of course, a good step to reproductive advantage. But so is being more attractive to or attracted to the opposite sex, shorter gestation, multiple births, and any number of other factors.
If an organism acquires, by chance, a feature that gives it and its offspring a reproductive advantage that results in even a tiny increase in net fertility rate, its offspring will prosper relative to those that don't have it. Over many generations the feature will become common.
The article addresses the issue of when a particular set of features differentiates organisms sufficiently that species separation occurs. As has been pointed out previously, what is at issue is the details of the mechanism, not its existence.
A theory, according to the OED, is:
You don't understand the concept of "fair and balanced." It means that for every person who expresses an accepted and scientifically justifiable opinion, you give equal or greater weight to selected whackos who disagree.
Then, once it becomes accepted that there is "no consensus" you split the difference, and find some even more extreme whackos to skew the "middle ground" even further. Eventually those with well-considered opinion are completely marginalized.
An honest scientist cannot win in this environment, because he or she is not willing to take ever-more-extreme positions to maintain "balance".
I've done a lot of mainframe development and a lot of Unix/Linux development; scarcely any Windows.
The main difference I see between mainframe development and *ix development culture is respect. With the mainframe you have to book time days in advance and work in the wee hours to make any changes. And you make damned sure that, when you're done, things work as they should.
With *ix development, things are laissez-faire. You send out a message a few hours/days/minutes in advance of some monumental change. Then you blame the users when they can't sign on to their system in the morning. Quote some recently-adopted standards if they argue.
Of course, I'm speaking of the early days of *ix. These systems are more and more critical, and the admins are trying to learn respect. But they're playing catch-up. There's nothing like the fear of taking down a $500/minute system to make you careful.
Windows development follows a similar pattern. The whole culture is so "personal computer" based that the concept of a year's continuous uptime is foreign.
The book describes magnetic core which was pretty well obsolete in 1971. It describes CPU, disk, and tape that have become smaller and faster, but that's about it. The monumental developments in programming languages - object-oriented, functional programming, concurrent programming - were in the past. Ditto for file systems, data structures, and operating systems.
If you learned computer architecture back then, you wouldn't have much difficulty with today. Not like the man-frozen-in-the-glacier movie scenario at all.
The development that wasn't even a gleam in the eye of computer scientists of the day is networking.
When you inflate a balloon you supply energy which is used to displace air in the atmosphere. There's no free lunch.
My point about EV was that, whatever mechanism you use, you'll need at least that much energy.
Suppose you rode your helium balloon to the moon and then cut your basket loose. By the time you fell to earth you'd have reached escape velocity. If you really could get to the moon with less energy you'd have invented a perpetual motion machine.