1) Large organizations typically use a fairly large amount of customized or completely custom software. These organizations pay people to do the work, and they trust them. The company I work for provides these services, and our clients pay us not only because we can do the work, but because over time they've found us trustworthy.
2) Much of the software being customized consists of off-the-shelf products. The problem he's worried about is not confined to OSS.
3) In certain industries (pharmaceuticals for example) that are regulated, software must go through a 'validation' process. This process ensures that the software meets the requirements and that strict configuration control is in place to prevent untested and/or unauthorized patches from making it into production. Circumventing validation would require a huge amount of collusion, and is therefore unlikely.
...Unfortunately, the model breaks down as soon as the core group involved in a project or distribution decides to corrupt the source, because they simply won't make the corrupted version public.
The problem with his argument is that the software in question is, generally, under the GPL. An organization large enough to suffer serious consequences through an introduced vulnerability (such as a national government) would draw attention to itself by using open source software, and inevitably someone would demand that they make their (corrupted) versions public. I don't buy that these corruptions would go unnoticed long enough to have serious consequences.
Never mind that his whole editorial is conjecture without proof that this has ever actually occurred (and no, bug list comparisons don't count. Sorry.)
I'm looking forward to trying out the new demo, and I'm interested in trying the vehicles out. Hopefully my 1.2 GHz Athlon w/GeForce 2 GTS can run the game at more than 5 fps:)
Personally, I'd like to see a PC version of Car Wars (old Steve Jackson game). Like Interstate 76 (rock on Daddy-O!:) but with the ability to hop out of the car and whoop some ass UT-style. There may be a game out there like that already, I dunno (haven't had a chance to play many games lately)... if so point me to it!
Check this out: the TypeMatrix EZR2030. They're not available yet (in production now) but I'm getting two right off the line. Same size as a laptop keyboard, same key action, thin and flat (in fact small enough to carry in your laptop bag), plus it has other neat features like putting the Enter and Backspace keys in the center of the keyboard where your stronger index fingers can use them. Oh, and it switches between Dvorak (which I use) and Qwerty. No wrist rest though - but you can get one of those separately.
Funny coincidence, today I just happened to be reading Fooled By Randomness, and came across a discourse almost exactly like this one. For example:
"...One conceivable way to discriminate between a scientific intellectual and a literary intellectual is by considering that a scientific intellectual can usually recognize the writing of another but that the literary intellectual would not be able to tell the difference between lines jotted down by a scientist and those by a glib non-scientist."
He goes on to reference other works including one by Sokal. Here's a useful (er, funny) item I found by reading the book: The Postmodernism Generator, based on the Dada Engine (a tool for generating random texts based on specified grammars).
This all reminds me of architecture school (shiver);)
I mean, they can do it on Star Trek, right? Artificial gravity would be the perfect tech for this purpose... putting a simulator capsule on a centrifuge is so, like, stone age.;)
I agree that learning facts is a secondary goal to learning skills, with the exception of some set of facts that everybody should know, as pointed out by other respondents. The *skills* a school should teach should include, IMHO, the following:
Communication - people need to be able to communicate articulately. This includes good spelling and grammar in my opinion, as they contribute to clear communication (/.ers may disagree;) I think this also includes reading comprehension.
Reasoning - this includes deductive analysis, recognition of logical fallacies, etc.
Study Skills - self-education
Social Skills
However, skills need to be practiced. I think any good school uses facts and knowledge as a medium to practice those skills, so they become reflexive; so in that sense they're necessary.
Nah, it just times out because my sucky implementation of Condorcet takes forever to compute;) I haven't had time to optimize it. I can write a slow application in any language - it has nothing to do with it running on.NET.
I should probably take the link out of my sig until I get it fixed though...
This article isn't worth discussing... there's no meat there, just a guy gloating over his one chance to respond to Mac zealots (hey fella, I sympathize, they're obnoxious).
I have to say I'm impressed with this NTFS hack, and it would be useful for me, but developing OSS ext2/ext3/reiserfs drivers for Windows would be equally useful. I use Explore2Fs and it's OK, but I'd rather have the support in Windows itself. This idea has never occurred to me, and it probably hasn't occurred to many others as well.
I sure don't need it. My wife and I only get broadcast analog, and I don't see our buying cable, a dish, or even HDTV anytime soon. Why pay for something that's really only a waste of my time? I've got more important and constructive things to do. Of course YMMV.
You're correct that election methods that involve ranking the candidates are a pain, especially for voters who aren't familiar with all the candidates. Some methods (like Condorcet) allow you to rank candidates you don't know much about similarly, perhaps at the bottom.
The vote count thing also is worth discussing. This is related to the "one person, one vote*" mantra that some people use against approval voting and other methods. The argument goes that with approval voting, some people will get more votes than others, so it's unfair. Here you're using the same argument to posit vote fraud.
However, look at what you can deduce from the information provided. With plurality voting, you can only vote for one candidate, so if you vote for candidate A, you can deduce nothing about the voter's feelings for candidate B other than he likes A more than B. So in reality you might as well put a question mark by B (and C, etc.)
With approval voting, the fact that you can vote for B whether you vote for A or not means that an unchecked box by B is equivalent to a 'no' vote. The voter actively disapproves of candidate B. When you look at it from this perspective, every voter gets the same representation, because there are no ambiguous votes. Detecting voting fraud is a matter of counting ballots, not votes.
You're correct though that it's not really possible to add up the votes themselves. Furthermore, it's not possible to report the election results in terms of absolute percentages; you can only report approval ratings. But I would argue that's more useful. It poses a problem however if some districts use approval while others use plurality, because you can't really compare the results, only the candidates chosen.
*"One person, one vote", by the way, grew out of the civil rights court cases of the fifties and sixties, and has more to do with congressional districting than election methods. It really means that each district should represent the same number of people, and doesn't imply that only one physical vote can be cast per person/election.
Also known as Instant Runoff Voting (IRV). This method has serious problems when examined according to technical fairness criteria.
The issue is that IRV works OK until a third party becomes viable - then, all bets are off. The article mentioned above quotes the following as an advantage:
It promotes a strong two-party system, ensuring stability in the parliamentary process.
Is this an advantage? I think not. The more common system, plurality or "first-past-the-post", which to be fair is even worse than IRV, does the same thing by artificially encouraging people to vote for front-runners. I would argue that any such artificial bias towards any party is a bad thing, and that the vote should reflect the true preferences of the voters as accurately as possible. IRV is an illusory fad in this regard.
Approval Voting and the Condorcet Method are much better. Condorcet is technically the best available method, but approval is (for the US anyway) also a good choice because it offers good technical compliance and ease of practical implementation.
Glad they've got a good voting machine. Next they should get a good election method. Instant Runoff Voting (IRV) has major problems (though there is a variant that attempts to correct them).
I'm a Georgia native, and though I'm a Democrat I generally agree with your assessment of the nature of Georgia Democrat-ness. Occam's Razor also suggests that your response is correct, and that conservatism here could explain the election*.
But it's hard to say. I know some people who disliked Barnes, some for his education policies (mainly teachers), some for the flag issue (though I know more who supported him), and some for other reasons. And I don't see any evidence in your post quoting results of specific polls that supported the election results.
None of this answers the question of whether the machines were tampered with, or whether the junk machines were sent to heavily Democratic precincts. There is no excuse for secret, unverified patches to voting machines, supplied by someone who heavily favors one party. Voting systems should be open-source, on reliable open hardware platforms, and the vulnerable points (which will exist) must be monitored by multiple parties.
This is a common (and I think valid) comment from Go players, and it's often mentioned to support the idea that Go is a "better" game. One highly-ranked Go player I knew was very snobby in this regard.
I think it's true that the search space for Go is much larger than for chess, but I suspect it's not so great as the Go players sometimes make it out to be. There are two factors in chess that provide variability that aren't really present in Go: first, in chess, the pieces are different. A knight on a given square is not the same as a rook on that square, whereas in Go a stone is a stone. Second, in chess the pieces can move repeatedly, and each move has multiple effects - it removes a piece from one location and adds it somewhere else. In Go, once you place a stone it stays there until it's captured.
Personally, for me the whole thing is moot, since I don't play either chess OR Go very well;) They're both sufficiently beyond my ability to master that to me they're equal in terms of difficulty. But I'd be curious to see a comparative study that takes the factors above into account.
Man-purse, whatever. I carry my Leatherman (micra) on my keychain. My cell phone goes on my belt, not because I like the way it looks there, but so I can keep it on 'vibrate' to avoid bothering my coworkers (hey, being nice pays off sometimes;). Everything else goes in my laptop bag.
I want to see the Dean campaign (or somebody!) release software implementing alternative voting methods. Everybody's talking about voting machines, but not many are talking about the software and algorithms, which make more of a difference IMHO. Dean has expressed support for Instant Runoff Voting (IRV), but there are better methods out there.
I've built an online poll demonstrating these methods (see my sig), but I wouldn't call my stuff release-worthy just yet (it's pretty slow, and I lost my source in a HD crash last week anyway)...
So what you are basically...heavy regulation of corporate accounting practices helped them.
I don't feel that all people are idiots, but some are (besides, knowledge is more important than intelligence in these things anyway IMHO). It seems obvious to me that a person can't take the time to research every product he might want to buy (not if she wants a life anyway). Think of the loss to the GNP if we were forced to do that. So we must delegate the task. Currently, there's Consumer Reports, a handful of smaller similar organizations, the media, and then there's the government. I don't feel that regulations inherently make anything better, especially if they're not enforced, and I'm not a fan of big government for its own sake, but I believe most regulations have arisen in response to abuses. I also believe that to the extent that people can trust the government to look out for them, the productivity gained outweighs the productivity lost to regulatory compliance. Which leads me to the next comment:
Do you know how much it costs to comply with government regulations?
As a matter of fact, I do. I'm currently working on an audit project for a large pharmaceutical company (FDA 21 CFR Part 11 related, in part, if you must know). And despite the impression you may have gotten from my earlier postings, everyone in that company I have worked with has expressed a real desire to do right by the law and by the customer. However, I also know that not all businesses (or let's say CEOs) are that professional, and many are perfectly willing to use dishonest, unfair or otherwise unethical business practices in the pursuit of a buck. There are a million examples in that vein (Microsoft, anyone?). Watchdogs are still needed.
Let me clarify my position a little bit. I believe regulation is a symptom of failure, and should be enacted as a last resort, but that in many cases it is necessary (IMHO the majority of current regs are necessary. Furthermore, I believe that most people are intelligent, and when presented with the facts and encouraged to think about their choices a little bit, will choose rationally.
To bring this discussion back on topic, I don't think we're there yet. Crucial to that vision is a free, strong and diverse media, so that people can make well-informed decisions, and I believe that this is one area where there can be no compromise. Loosening restrictions on the media will result in more regulation in the end, not less.
Slashdot Mirror
Other issues:
1) Large organizations typically use a fairly large amount of customized or completely custom software. These organizations pay people to do the work, and they trust them. The company I work for provides these services, and our clients pay us not only because we can do the work, but because over time they've found us trustworthy.
2) Much of the software being customized consists of off-the-shelf products. The problem he's worried about is not confined to OSS.
3) In certain industries (pharmaceuticals for example) that are regulated, software must go through a 'validation' process. This process ensures that the software meets the requirements and that strict configuration control is in place to prevent untested and/or unauthorized patches from making it into production. Circumventing validation would require a huge amount of collusion, and is therefore unlikely.
...Unfortunately, the model breaks down as soon as the core group involved in a project or distribution decides to corrupt the source, because they simply won't make the corrupted version public.
The problem with his argument is that the software in question is, generally, under the GPL. An organization large enough to suffer serious consequences through an introduced vulnerability (such as a national government) would draw attention to itself by using open source software, and inevitably someone would demand that they make their (corrupted) versions public. I don't buy that these corruptions would go unnoticed long enough to have serious consequences.
Never mind that his whole editorial is conjecture without proof that this has ever actually occurred (and no, bug list comparisons don't count. Sorry.)
I'm looking forward to trying out the new demo, and I'm interested in trying the vehicles out. Hopefully my 1.2 GHz Athlon w/GeForce 2 GTS can run the game at more than 5 fps :)
:) but with the ability to hop out of the car and whoop some ass UT-style. There may be a game out there like that already, I dunno (haven't had a chance to play many games lately)... if so point me to it!
Personally, I'd like to see a PC version of Car Wars (old Steve Jackson game). Like Interstate 76 (rock on Daddy-O!
Check this out: the TypeMatrix EZR2030. They're not available yet (in production now) but I'm getting two right off the line. Same size as a laptop keyboard, same key action, thin and flat (in fact small enough to carry in your laptop bag), plus it has other neat features like putting the Enter and Backspace keys in the center of the keyboard where your stronger index fingers can use them. Oh, and it switches between Dvorak (which I use) and Qwerty. No wrist rest though - but you can get one of those separately.
"...One conceivable way to discriminate between a scientific intellectual and a literary intellectual is by considering that a scientific intellectual can usually recognize the writing of another but that the literary intellectual would not be able to tell the difference between lines jotted down by a scientist and those by a glib non-scientist."
He goes on to reference other works including one by Sokal. Here's a useful (er, funny) item I found by reading the book: The Postmodernism Generator , based on the Dada Engine (a tool for generating random texts based on specified grammars).
This all reminds me of architecture school (shiver)
I mean, they can do it on Star Trek, right? Artificial gravity would be the perfect tech for this purpose... putting a simulator capsule on a centrifuge is so, like, stone age. ;)
However, skills need to be practiced. I think any good school uses facts and knowledge as a medium to practice those skills, so they become reflexive; so in that sense they're necessary.
Nah, it just times out because my sucky implementation of Condorcet takes forever to compute ;) I haven't had time to optimize it. I can write a slow application in any language - it has nothing to do with it running on .NET.
I should probably take the link out of my sig until I get it fixed though...
This article isn't worth discussing... there's no meat there, just a guy gloating over his one chance to respond to Mac zealots (hey fella, I sympathize, they're obnoxious).
Not worth having a real discussion over though.
I have to say I'm impressed with this NTFS hack, and it would be useful for me, but developing OSS ext2/ext3/reiserfs drivers for Windows would be equally useful. I use Explore2Fs and it's OK, but I'd rather have the support in Windows itself. This idea has never occurred to me, and it probably hasn't occurred to many others as well.
Wish I had mod points so your response would get the credit it deserves, and so the arrogant moron you're responding to would get modded down.
From the article:
Similar designs use ethanol (beverage alcohol) but currently have substantially lower efficiencies
So theoretically I could power my laptop with beer? Hmmm, which to choose: code, or drink.... code, or drink.... dammit!
Or if I want to play Quake instead of coding, maybe I should use something stronger, like, say, vodka. That'd be cool.
I sure don't need it. My wife and I only get broadcast analog, and I don't see our buying cable, a dish, or even HDTV anytime soon. Why pay for something that's really only a waste of my time? I've got more important and constructive things to do. Of course YMMV.
You're correct that election methods that involve ranking the candidates are a pain, especially for voters who aren't familiar with all the candidates. Some methods (like Condorcet) allow you to rank candidates you don't know much about similarly, perhaps at the bottom.
The vote count thing also is worth discussing. This is related to the "one person, one vote*" mantra that some people use against approval voting and other methods. The argument goes that with approval voting, some people will get more votes than others, so it's unfair. Here you're using the same argument to posit vote fraud.
However, look at what you can deduce from the information provided. With plurality voting, you can only vote for one candidate, so if you vote for candidate A, you can deduce nothing about the voter's feelings for candidate B other than he likes A more than B. So in reality you might as well put a question mark by B (and C, etc.)
With approval voting, the fact that you can vote for B whether you vote for A or not means that an unchecked box by B is equivalent to a 'no' vote. The voter actively disapproves of candidate B. When you look at it from this perspective, every voter gets the same representation, because there are no ambiguous votes. Detecting voting fraud is a matter of counting ballots, not votes.
You're correct though that it's not really possible to add up the votes themselves. Furthermore, it's not possible to report the election results in terms of absolute percentages; you can only report approval ratings. But I would argue that's more useful. It poses a problem however if some districts use approval while others use plurality, because you can't really compare the results, only the candidates chosen.
*"One person, one vote", by the way, grew out of the civil rights court cases of the fifties and sixties, and has more to do with congressional districting than election methods. It really means that each district should represent the same number of people, and doesn't imply that only one physical vote can be cast per person/election.
Doh! Typing from memory ;)
Also known as Instant Runoff Voting (IRV). This method has serious problems when examined according to technical fairness criteria.
The issue is that IRV works OK until a third party becomes viable - then, all bets are off. The article mentioned above quotes the following as an advantage:
It promotes a strong two-party system, ensuring stability in the parliamentary process.
Is this an advantage? I think not. The more common system, plurality or "first-past-the-post", which to be fair is even worse than IRV, does the same thing by artificially encouraging people to vote for front-runners. I would argue that any such artificial bias towards any party is a bad thing, and that the vote should reflect the true preferences of the voters as accurately as possible. IRV is an illusory fad in this regard.
Approval Voting and the Condorcet Method are much better. Condorcet is technically the best available method, but approval is (for the US anyway) also a good choice because it offers good technical compliance and ease of practical implementation.
Glad they've got a good voting machine. Next they should get a good election method. Instant Runoff Voting (IRV) has major problems (though there is a variant that attempts to correct them).
I think they should be just as worried that people might mistake their real news ticker for cartoon news...
Oh wait. Guess there's no chance of that happening... now I'm depressed.
I'm a Georgia native, and though I'm a Democrat I generally agree with your assessment of the nature of Georgia Democrat-ness. Occam's Razor also suggests that your response is correct, and that conservatism here could explain the election*.
But it's hard to say. I know some people who disliked Barnes, some for his education policies (mainly teachers), some for the flag issue (though I know more who supported him), and some for other reasons. And I don't see any evidence in your post quoting results of specific polls that supported the election results.
None of this answers the question of whether the machines were tampered with, or whether the junk machines were sent to heavily Democratic precincts. There is no excuse for secret, unverified patches to voting machines, supplied by someone who heavily favors one party. Voting systems should be open-source, on reliable open hardware platforms, and the vulnerable points (which will exist) must be monitored by multiple parties.
*Zell Miller is a closet Republican.
This is a common (and I think valid) comment from Go players, and it's often mentioned to support the idea that Go is a "better" game. One highly-ranked Go player I knew was very snobby in this regard.
;) They're both sufficiently beyond my ability to master that to me they're equal in terms of difficulty. But I'd be curious to see a comparative study that takes the factors above into account.
I think it's true that the search space for Go is much larger than for chess, but I suspect it's not so great as the Go players sometimes make it out to be. There are two factors in chess that provide variability that aren't really present in Go: first, in chess, the pieces are different. A knight on a given square is not the same as a rook on that square, whereas in Go a stone is a stone. Second, in chess the pieces can move repeatedly, and each move has multiple effects - it removes a piece from one location and adds it somewhere else. In Go, once you place a stone it stays there until it's captured.
Personally, for me the whole thing is moot, since I don't play either chess OR Go very well
Man-purse, whatever. I carry my Leatherman (micra) on my keychain. My cell phone goes on my belt, not because I like the way it looks there, but so I can keep it on 'vibrate' to avoid bothering my coworkers (hey, being nice pays off sometimes ;). Everything else goes in my laptop bag.
Me: Computer! What's happening in Iraq today?
Computer: Three mo'e South Car'linan soldiers were killed today in Iraq as attacks intensified, cuss it all t' tarnation.
I want to see the Dean campaign (or somebody!) release software implementing alternative voting methods. Everybody's talking about voting machines, but not many are talking about the software and algorithms, which make more of a difference IMHO. Dean has expressed support for Instant Runoff Voting (IRV), but there are better methods out there.
I've built an online poll demonstrating these methods (see my sig), but I wouldn't call my stuff release-worthy just yet (it's pretty slow, and I lost my source in a HD crash last week anyway)...
The only problem is that opening the door may cause the game to start/stop ;)
I don't feel that all people are idiots, but some are (besides, knowledge is more important than intelligence in these things anyway IMHO). It seems obvious to me that a person can't take the time to research every product he might want to buy (not if she wants a life anyway). Think of the loss to the GNP if we were forced to do that. So we must delegate the task. Currently, there's Consumer Reports, a handful of smaller similar organizations, the media, and then there's the government. I don't feel that regulations inherently make anything better, especially if they're not enforced, and I'm not a fan of big government for its own sake, but I believe most regulations have arisen in response to abuses. I also believe that to the extent that people can trust the government to look out for them, the productivity gained outweighs the productivity lost to regulatory compliance. Which leads me to the next comment:
Do you know how much it costs to comply with government regulations?
As a matter of fact, I do. I'm currently working on an audit project for a large pharmaceutical company (FDA 21 CFR Part 11 related, in part, if you must know). And despite the impression you may have gotten from my earlier postings, everyone in that company I have worked with has expressed a real desire to do right by the law and by the customer. However, I also know that not all businesses (or let's say CEOs) are that professional, and many are perfectly willing to use dishonest, unfair or otherwise unethical business practices in the pursuit of a buck. There are a million examples in that vein (Microsoft, anyone?). Watchdogs are still needed.
Let me clarify my position a little bit. I believe regulation is a symptom of failure, and should be enacted as a last resort, but that in many cases it is necessary (IMHO the majority of current regs are necessary. Furthermore, I believe that most people are intelligent, and when presented with the facts and encouraged to think about their choices a little bit, will choose rationally.
To bring this discussion back on topic, I don't think we're there yet. Crucial to that vision is a free, strong and diverse media, so that people can make well-informed decisions, and I believe that this is one area where there can be no compromise. Loosening restrictions on the media will result in more regulation in the end, not less.