Sorry, I've also tried to explain it in earlier post(s). There are no "real terms". And I am not mind reading. Indeed, different people will map the same word to different numbers and the same number mean different words to different people. But words are no more canonical than numbers. They are both languages only. In a lot of competitions, judges give a 0.0-10.0 score and scores are summed up / averaged to get the final score for each competitors. Those competitions are done so because the competitors' performance are fundamentally subjective. Similar to politicians. There are NEVER valid objective measures to gauge a leader or a representative. If a candidate is hated by a significant group of voters sincerely, the ability to vote him down without hindering the expression of their own favorites is an advantage of score voting, NOT disadvantage. This ability reduce the risk of vote splitting and dissolve the tactic of threatening used by more wealthy or popular candidates. In rank based voting, a candidate whom the majority strongly disagree with may still win if the opposing forces are too segmented and don't have a strong leader. Then pretty soon we went back to racing for popularity, rather than who's really better.
As long as voters are educated of the possibility that others may bullet vote and approval vote (score the favorite(s) 99 and all others 0), everything will be fine. At least score voting is more transparent and easier to recount than IRV or any Condorcet. And the election process/result is a lot more stable and predictable. If you still feel that scoring is not objective enough while your ranking is objective, well, it may be true in the sense that ranking is objectively skewed and objectively unreliable, as shown by Arrow himself. Remember, IRV violates monotonicity, thus manipulable by propaganda machine far easier, whereas score voting promises everyone can still cast a sincere vote safely, no matter what tactics or "holes" involved.
Give me an example of "hole" in score voting that involves insincere votes. Then they are really holes. Else they are just clever voting. And each group can do clever voting themselves. It's fair. Unless you say there's a group more clever / stupid than others... More power to clever people, ok?
Now here is where our fundamental disagreement stems from. I see B being the correct outcome while you don't. (Actually I don't even encourage any "normalization" for single-seat elections.)
We are about electing one single representative for a whole group of people here. So what he/she say and do ought to balance opinions of all voters. Your legacy definition of "favorite" lingered from FPTP era ignores the needs of minorities. Tyranny of majority is dangerous. Britain and America seems not suffering from that only because they don't have a different enough political minority currently. Look at those African nations. Look at those civic wars. From election we want to find out the one who is supported the most by all people, and this does not necessarily equal to the one who is supported the most by the most people. And your example nicely illustrate how the "strength of preference" I mentioned kicks in. Only score voting can let voters express the satisfactory level deviation among candidates. Arrow's ranking can only rank them, and assumes all A > B > C means the same. The impossibility theorem is a result of such information hiding, forcing the vote counting mechanisms to guess and introduce unsatisfactory results from time to time.
(If we consider multi-seat elections then that's another matter. There would be no longer need to find someone who can balance everybody. We only need proportional representation.)
You said you want to normalize the figures ("Since there is no way to normalize these figures in a standard value system...") and I provide a method. I never said normalization should be done by default. Ranking may be less ambiguous than scoring to you, but what makes you so sure it reflect voters' wishes more accurately? See my another reply to your post.
And No, most system don't give people a chances to express how much they care about the election, and how different they see their 1st and last choices are. They can only indicate that indirectly by either not go to vote, or cast a blank vote. They cannot express that they feel all candidates are bad, and tell the official there are someone they thought to be least evil at the same time. Most systems can only do a "yay" or "nay" in this area, either you have a preference and you come to vote, or you don't have a preference and you don't come to vote.
There are many possible systems that give people more choices than merely "yay" or "nay". The problem with using an arbitrary numeric scale is that you have no way of knowing what different numbers may mean to different people. Somebody who "kind of likes" candidate X may give him 67 points (about 2/3), while someone else who feels the same way might give that same candidate 55 points (feeling that more than half is good), and yet another person who also feels the same about the candidate might vote 80 points.
Because of this uncertainty in the scale, simply adding up all the points for the candidates is not a valid measure of who was actually the preferred person. I think it would be pretty easy to show that it would be possible to elect a candidate who was not the actual favorite. Not common, to be sure, but possible.
You see your example a display of how people can't express "kind of like" objectively and equally; I see your example a display people aren't thinking the same when they say "kind of like", thus it is the concept "kind of like" that are too subjective, not the inverse.
Arrow assumes ranking the basis of any preferences, but it is very likely that when 2 people vote somebody as their 2nd choice, the meaning behind is fundamentally different. Alice and Bob both vote A > B > C in Condorcet. But in Approval, maybe Alice would like to vote A only but Bob A + B. Ranking hides that. Scoring expresses that. Alice can vote B a 30 and Bob can vote B a 70. As long as a voter has one single scale to judge all candidates, the combined result will be objective relative to who's voting.
As long as people know how the score they gave are used, it is their own responsibility to translate their feeling correctly to numbers. Ranking may look more manageable and objective to you, but it is only because ranking does a lossy compression to people's opinion and hides the dirty stuff. And then they will be surprised, when the election result does not match their mental diagram occasionally.
That does "normalize", in the sense of scaling all values to the range of 1 to 100 (or 0 to 100, whatever). But in an election of any size, someone is bound to vote 1 and someone else vote 100, so that no scaling would actually be done.
Maybe there's some misunderstanding here. What I mean is to scale each one's vote individually. That is, if I vote A the lowest and B the highest, my vote will be treated as A scored 0 and B scored 99. Then the same algorithm applies to your vote.
As long as everyone understand how their votes will be accounted and what effect will be done by how they vote, the votes can be objectively compared.
Also, there is a simple way to do renormalization if you really want to. Treat the lowest scored candidate as scored 0 and the highest scored 100 and scale the others as such, for each individual vote. Normalization finished.
Many people don't vote because they don't think the candidates are that much different. Range voting/score voting (choose a name you like) let people express how strong their preferences on the election are. They will have more choices than vote or absent.
When a spec/standard is not followed or even wilfully violated by major implementations, could we really see the spec as real? When the spec writers don't have content writers in mind, what good is being "neutral"? You mentioned Nokia and Siemens. See? They don't write webpages for a living. People don't rely on them to read webpages. They are out of touch to the WWW. Therefore they wasted all the time on XHTML, web engines produce a bunch of <p/> <br/> to their flavour, and browsers strip those slashes back out when interpreting.
IE were unchallenged for years because new competitors need to be compatible to webpages that relies on IE's out-of-spec behaviours. At least HTML5 promises there will not be any situations left unspecified.
It's because those "companies" behind those civilian aircraft designs are all somewhat nation based. So they are inherently bureaucratic, like government departments.
Could someone fluent in Mandarin comment on the translation? I'm especially interested in this translation: "The goal of the project is to boost our economy, not theirs." Is the implication that the Chinese gov't wants their trading partners to suffer, or that their partners' situation just isn't important?
China's economy benefits when their trading partners economies benefit. If it's the former (or even the latter, to a degree), that suggests they have other priorities in mind.
Although my mother tongue is Cantonese, not Mandarin, such line is indeed there in the original article. It's in the 3rd paragraph after the section subtitle "Door close or open?", quoted as a thought from the higher-up. It didn't really said of "goal" but the meaning is mostly accurate, just that the "not theirs" emphasis were not there.
However, I must say the translation posted here is a little bit twisted, since I am quite sure the implication of that line is not about making the trading partners suffer but only about making themselves the more benefited in the trading process.
How about really spend some time on teaching HTML and CSS, put them as the main characters of the class and start on Javascript later, maybe next semester?
I think most of us who come here know this already. It's an almost-established fact among this field. My concern on this interview is, who else will also read this? Are there anybody that previously don't know that will gain this knowledge? Are there any effect on them? I have not much confidence that those people in power may read Information Economics and Policy and rethink what they are doing right now seriously.
Your idea remind me of magnet link in Bittorrent. Although they have significant difference, but the outlook is very similiar: a link with hash!
One rub: since computer is always advancing, brute-force cracking become easier and easier, which in return asks for key lengthening, or at least key switching, every n years. Are we going to be required to "renew" our bookmarks every n years too? How to handle the "hyperlink expiration"?
We have Octopus card here in HK, something that's closest to electronic cash than probably every other things on earth. It is a smart card that could have no identity attached. A lot of money transaction like public transit, fast-food restaurant, supermarket, 7-11, etc. can run through it. Spend the money stored inside, recharge, and spend. All those little cents are slowly being abandoned by people here now.
Somebody here in Hong Kong made fun that the Peace Prize is very fitting to the mainland China government motto, "Peace Rise".
Originally, CCTV produced a TV program called "Rise of Great Powers". Feeling bad of being finger pointed, the government soon changed the slogan to "Rise in Peace" (something like that). Now, in the meantime they discussing how to "rise peacefully", they got a citizen born in China and live in China being given a Peace Prize. They should "celebrate" their everlasting effort now got international recognition!
Back on topic. Media here in Hong Kong succeeded to maintain a phone call to Liu Xia on that day, notified her the latest information, and asked her for some comment. And she spoke "I am proud of him. He is innocent" something like that. No wonder now the police block the phone line too. Seriously speaking, there are messages on Twitter saying that some related people were detented when they wanted to have a good dinner together for a private celebration. (Hope things like that won't be copied to Hong Kong.)
Now that's a plausible explanation of the transparent fuselage idea. Just like usually people won't want a transparent bus roof, a transparent roof for the plane is also not that good an idea to passengers. However, some transparency for the roof of aisles may not be bad, especially for delta wing aircrafts.
I think what this idea mostly ends up is just bigger windows for planes, like those of air-conditioned buses.
As I said, Cantonese aligns better against old poems writings, especially those written in the more "glorious" eras of China. It even get its pronunciation acceptably matching from the old dictionaries written at that time. So one of the forces of rebound from the speaking language unification policy also stems from nationalism. A few people in Canton even proclaim in forums that Cantonese should better be the standard instead. Some people also use Sun Yat-sen, the Father of Republic of China, being born in Canton as a spiritual support for their claims and actions. Both sides are somehow using roughly the same nationalism flag. That's what smells...
I live in Hong Kong, a rare place that is remotely ruled by the Communist Party but not shadowed by the GFW (yet).
You reminded me of the twitter mock in China just blocked JiangNanXi in their search function. Just today people in Canton (or GuangDong province you may call) assembled together in JiangNanXi Road protesting for the attempts of elimination of Cantonese from Canton. The government first suggested an elimination of Cantonese from local TV programmes and then a famous general born in Canton several hundred years ago had his punch line removed from his statue in a park. Meanwhile some elementary schools are found to be prohibiting students from speaking Cantonese since quite some times ago. Youngsters start to feel their own culture and traditions threatened.
What make things more silly is, it is a common knowledge among linguists that Cantonese, compared to Mandarin, is more matching in rhyme for reading old poems written in Tang and Song Dynasties (more than 1000 years ago). And while the government always calls Cantonese a dialect, Cantonese and Mandarin are not mutually intelligible, sometimes even difference in grammar. (Those pro-government bodies even use the lack of mutual intelligibility as a reason to further push their Mandarin, to "accommodate" for people from other provinces.) You should smell something wrong now...
Slashdot Mirror
Sorry, I've also tried to explain it in earlier post(s). There are no "real terms". And I am not mind reading. Indeed, different people will map the same word to different numbers and the same number mean different words to different people. But words are no more canonical than numbers. They are both languages only. In a lot of competitions, judges give a 0.0-10.0 score and scores are summed up / averaged to get the final score for each competitors. Those competitions are done so because the competitors' performance are fundamentally subjective. Similar to politicians. There are NEVER valid objective measures to gauge a leader or a representative. If a candidate is hated by a significant group of voters sincerely, the ability to vote him down without hindering the expression of their own favorites is an advantage of score voting, NOT disadvantage. This ability reduce the risk of vote splitting and dissolve the tactic of threatening used by more wealthy or popular candidates. In rank based voting, a candidate whom the majority strongly disagree with may still win if the opposing forces are too segmented and don't have a strong leader. Then pretty soon we went back to racing for popularity, rather than who's really better.
As long as voters are educated of the possibility that others may bullet vote and approval vote (score the favorite(s) 99 and all others 0), everything will be fine. At least score voting is more transparent and easier to recount than IRV or any Condorcet. And the election process/result is a lot more stable and predictable. If you still feel that scoring is not objective enough while your ranking is objective, well, it may be true in the sense that ranking is objectively skewed and objectively unreliable, as shown by Arrow himself. Remember, IRV violates monotonicity, thus manipulable by propaganda machine far easier, whereas score voting promises everyone can still cast a sincere vote safely, no matter what tactics or "holes" involved.
Give me an example of "hole" in score voting that involves insincere votes. Then they are really holes. Else they are just clever voting. And each group can do clever voting themselves. It's fair. Unless you say there's a group more clever / stupid than others... More power to clever people, ok?
Now here is where our fundamental disagreement stems from. I see B being the correct outcome while you don't. (Actually I don't even encourage any "normalization" for single-seat elections.)
We are about electing one single representative for a whole group of people here. So what he/she say and do ought to balance opinions of all voters. Your legacy definition of "favorite" lingered from FPTP era ignores the needs of minorities. Tyranny of majority is dangerous. Britain and America seems not suffering from that only because they don't have a different enough political minority currently. Look at those African nations. Look at those civic wars. From election we want to find out the one who is supported the most by all people, and this does not necessarily equal to the one who is supported the most by the most people. And your example nicely illustrate how the "strength of preference" I mentioned kicks in. Only score voting can let voters express the satisfactory level deviation among candidates. Arrow's ranking can only rank them, and assumes all A > B > C means the same. The impossibility theorem is a result of such information hiding, forcing the vote counting mechanisms to guess and introduce unsatisfactory results from time to time.
(If we consider multi-seat elections then that's another matter. There would be no longer need to find someone who can balance everybody. We only need proportional representation.)
You said you want to normalize the figures ("Since there is no way to normalize these figures in a standard value system...") and I provide a method. I never said normalization should be done by default. Ranking may be less ambiguous than scoring to you, but what makes you so sure it reflect voters' wishes more accurately? See my another reply to your post.
And No, most system don't give people a chances to express how much they care about the election, and how different they see their 1st and last choices are. They can only indicate that indirectly by either not go to vote, or cast a blank vote. They cannot express that they feel all candidates are bad, and tell the official there are someone they thought to be least evil at the same time. Most systems can only do a "yay" or "nay" in this area, either you have a preference and you come to vote, or you don't have a preference and you don't come to vote.
There are many possible systems that give people more choices than merely "yay" or "nay". The problem with using an arbitrary numeric scale is that you have no way of knowing what different numbers may mean to different people. Somebody who "kind of likes" candidate X may give him 67 points (about 2/3), while someone else who feels the same way might give that same candidate 55 points (feeling that more than half is good), and yet another person who also feels the same about the candidate might vote 80 points.
Because of this uncertainty in the scale, simply adding up all the points for the candidates is not a valid measure of who was actually the preferred person. I think it would be pretty easy to show that it would be possible to elect a candidate who was not the actual favorite. Not common, to be sure, but possible.
You see your example a display of how people can't express "kind of like" objectively and equally; I see your example a display people aren't thinking the same when they say "kind of like", thus it is the concept "kind of like" that are too subjective, not the inverse.
Arrow assumes ranking the basis of any preferences, but it is very likely that when 2 people vote somebody as their 2nd choice, the meaning behind is fundamentally different. Alice and Bob both vote A > B > C in Condorcet. But in Approval, maybe Alice would like to vote A only but Bob A + B. Ranking hides that. Scoring expresses that. Alice can vote B a 30 and Bob can vote B a 70. As long as a voter has one single scale to judge all candidates, the combined result will be objective relative to who's voting.
As long as people know how the score they gave are used, it is their own responsibility to translate their feeling correctly to numbers. Ranking may look more manageable and objective to you, but it is only because ranking does a lossy compression to people's opinion and hides the dirty stuff. And then they will be surprised, when the election result does not match their mental diagram occasionally.
That does "normalize", in the sense of scaling all values to the range of 1 to 100 (or 0 to 100, whatever). But in an election of any size, someone is bound to vote 1 and someone else vote 100, so that no scaling would actually be done.
Maybe there's some misunderstanding here. What I mean is to scale each one's vote individually. That is, if I vote A the lowest and B the highest, my vote will be treated as A scored 0 and B scored 99. Then the same algorithm applies to your vote.
As long as everyone understand how their votes will be accounted and what effect will be done by how they vote, the votes can be objectively compared.
Also, there is a simple way to do renormalization if you really want to. Treat the lowest scored candidate as scored 0 and the highest scored 100 and scale the others as such, for each individual vote. Normalization finished.
I disagree with what you say.
Many people don't vote because they don't think the candidates are that much different. Range voting/score voting (choose a name you like) let people express how strong their preferences on the election are. They will have more choices than vote or absent.
Just come to me: How about deriving from light pressure? It may be doable...
Is it my delusion or the font size is really smaller than in the past? Maybe the fault of those white space...
Nothing prevent a jet to be also a bi-plane.
When a spec/standard is not followed or even wilfully violated by major implementations, could we really see the spec as real? When the spec writers don't have content writers in mind, what good is being "neutral"? You mentioned Nokia and Siemens. See? They don't write webpages for a living. People don't rely on them to read webpages. They are out of touch to the WWW. Therefore they wasted all the time on XHTML, web engines produce a bunch of <p /> <br /> to their flavour, and browsers strip those slashes back out when interpreting.
IE were unchallenged for years because new competitors need to be compatible to webpages that relies on IE's out-of-spec behaviours. At least HTML5 promises there will not be any situations left unspecified.
It's because those "companies" behind those civilian aircraft designs are all somewhat nation based. So they are inherently bureaucratic, like government departments.
You are assuming the system can and will saturate 30 cat5 cables, which may or may not be true.
Agree.
It's a serious failure that we still can't talk with dolphins, albeit we know they do have some form of languages. What have those biologist done?
Could someone fluent in Mandarin comment on the translation? I'm especially interested in this translation: "The goal of the project is to boost our economy, not theirs." Is the implication that the Chinese gov't wants their trading partners to suffer, or that their partners' situation just isn't important?
China's economy benefits when their trading partners economies benefit. If it's the former (or even the latter, to a degree), that suggests they have other priorities in mind.
Although my mother tongue is Cantonese, not Mandarin, such line is indeed there in the original article. It's in the 3rd paragraph after the section subtitle "Door close or open?", quoted as a thought from the higher-up. It didn't really said of "goal" but the meaning is mostly accurate, just that the "not theirs" emphasis were not there.
However, I must say the translation posted here is a little bit twisted, since I am quite sure the implication of that line is not about making the trading partners suffer but only about making themselves the more benefited in the trading process.
How about really spend some time on teaching HTML and CSS, put them as the main characters of the class and start on Javascript later, maybe next semester?
Just like how food prices rise, and farmers gain nothing.
Well, in fact, a press do provides editors to book authors, so it does contribute. It's just a matter of how much the worth.
I think most of us who come here know this already. It's an almost-established fact among this field. My concern on this interview is, who else will also read this? Are there anybody that previously don't know that will gain this knowledge? Are there any effect on them? I have not much confidence that those people in power may read Information Economics and Policy and rethink what they are doing right now seriously.
Your idea remind me of magnet link in Bittorrent. Although they have significant difference, but the outlook is very similiar: a link with hash!
One rub: since computer is always advancing, brute-force cracking become easier and easier, which in return asks for key lengthening, or at least key switching, every n years. Are we going to be required to "renew" our bookmarks every n years too? How to handle the "hyperlink expiration"?
We have Octopus card here in HK, something that's closest to electronic cash than probably every other things on earth. It is a smart card that could have no identity attached. A lot of money transaction like public transit, fast-food restaurant, supermarket, 7-11, etc. can run through it. Spend the money stored inside, recharge, and spend. All those little cents are slowly being abandoned by people here now.
Somebody here in Hong Kong made fun that the Peace Prize is very fitting to the mainland China government motto, "Peace Rise".
Originally, CCTV produced a TV program called "Rise of Great Powers". Feeling bad of being finger pointed, the government soon changed the slogan to "Rise in Peace" (something like that). Now, in the meantime they discussing how to "rise peacefully", they got a citizen born in China and live in China being given a Peace Prize. They should "celebrate" their everlasting effort now got international recognition!
Back on topic. Media here in Hong Kong succeeded to maintain a phone call to Liu Xia on that day, notified her the latest information, and asked her for some comment. And she spoke "I am proud of him. He is innocent" something like that. No wonder now the police block the phone line too. Seriously speaking, there are messages on Twitter saying that some related people were detented when they wanted to have a good dinner together for a private celebration. (Hope things like that won't be copied to Hong Kong.)
What a wonderful "peace rise".
Now that's a plausible explanation of the transparent fuselage idea. Just like usually people won't want a transparent bus roof, a transparent roof for the plane is also not that good an idea to passengers. However, some transparency for the roof of aisles may not be bad, especially for delta wing aircrafts.
I think what this idea mostly ends up is just bigger windows for planes, like those of air-conditioned buses.
As I said, Cantonese aligns better against old poems writings, especially those written in the more "glorious" eras of China. It even get its pronunciation acceptably matching from the old dictionaries written at that time. So one of the forces of rebound from the speaking language unification policy also stems from nationalism. A few people in Canton even proclaim in forums that Cantonese should better be the standard instead. Some people also use Sun Yat-sen, the Father of Republic of China, being born in Canton as a spiritual support for their claims and actions. Both sides are somehow using roughly the same nationalism flag. That's what smells...
I live in Hong Kong, a rare place that is remotely ruled by the Communist Party but not shadowed by the GFW (yet).
You reminded me of the twitter mock in China just blocked JiangNanXi in their search function. Just today people in Canton (or GuangDong province you may call) assembled together in JiangNanXi Road protesting for the attempts of elimination of Cantonese from Canton. The government first suggested an elimination of Cantonese from local TV programmes and then a famous general born in Canton several hundred years ago had his punch line removed from his statue in a park. Meanwhile some elementary schools are found to be prohibiting students from speaking Cantonese since quite some times ago. Youngsters start to feel their own culture and traditions threatened.
What make things more silly is, it is a common knowledge among linguists that Cantonese, compared to Mandarin, is more matching in rhyme for reading old poems written in Tang and Song Dynasties (more than 1000 years ago). And while the government always calls Cantonese a dialect, Cantonese and Mandarin are not mutually intelligible, sometimes even difference in grammar. (Those pro-government bodies even use the lack of mutual intelligibility as a reason to further push their Mandarin, to "accommodate" for people from other provinces.) You should smell something wrong now...