We go thru this all the time with them, they push prices up to where they get worried we might actually go find an alternative, then bring it down just enough ( but higher then before ) to quiet us down and lose interest in alternatives. Its a cycle that most people are too stupid to see, and thus we are stuck in it.
You're probably right, but lets hope people are a little smarter this time. Maybe this time we have enough momentum and investment in alternative energy research to make the changes. The worst thing for alternative energy projects is that their return on investment calculations are so uncertain because OPEC can just drop oil prices and wipe the project out. What we really need is a higher petroleum tax to offset that effect. Not only will it cover that problem, but it could pay for the alternative energy research and implementation in addition to covering the cost of the externalities (pollution, etc) of the burning of fossil fuels so at least the market is more fair. But of course politicians and people's short sightedness are in the mix, so again, you're probably right. Of course the other possibility is that they really are running out of cheap oil and they aren't bluffing much. We all know it's going to happen sooner or later, it's just the when question.
Seems like a pretty weak test. Or at least not what I was thinking of from the question and lead in reference to Neo.
Experts and extraordinarily talented people in a specific domain do often report the feeling of everything else moving in slow motion. When you're so good at something or have practiced it so many times, you can read the nuances of a situation and react to them so quickly that it's comparable to if everything else had actually slowed down. I've experienced it in a limited way a few times, enough to know what people are talking about when they refer to it.
I can't believe they haven't figured out this is their biggest problem. Not having a specified license after a year means that they are going to be in a license mishmash for years to come. They've already copied Wikipedia articles, so those have to be GFDL. If they go with another license, either they need to get rid of all their GFDL material (a step back) or go forward with multiple licenses and all the incompatibility mess that goes along with that. I for one refuse to contribute until they get the license mess sorted out. And I like the idea of a project where I don't have to put up with all the BS that comes with Wikipedia.
Is trivia not information? - Of course it's information, but that doesn't mean it makes a better article. I submit that the highest quality article is one that is tightly focused on the topic, covers the most important information about the topic, and allocates space in relation to the information's importance to the topic. That way the reader can tell what is important and comes away from an article getting the most important information about the subject. That has the greatest utility to someone coming to use the article for it's information. Yes Wiki is not paper, but that doesn't mean there is space for 40mb of information about a topic all in one article. The less and less important information should be moved to sub articles that are focused on that aspect of a topic. That can go all the way down to the limit of what is verifiable information in credible sources. If that's done properly all the verifiable information is there, it's just organized properly.
The opposite of what I'm getting at is an article with no prioritization of material, random unimportant bits mixed with important ones. That's what trivia bits in articles amount to. Any trivia in an article means more important material has been skipped to give space to the trivia and that results in a lower quality article. Certainly no one is stopping anyone from creating triviapedia where that type of thing would be welcome, but trying to cram trivia in to lower the quality of articles intended to be high quality presentations of information is the problem. Just put it elsewhere, somewhere that's designed for it and there is no problem.
I also submit that what I have proposed here (and is what results when you distill down all of the policy and guideline articles about how to write a great Wikipedia articles) results in self evidently higher quality articles. Again the only conflict comes in trying to use something for something it's not and wondering why there's a conflict.
Exactly. And I'm sure a lot of other people feel the same way. This could make Redhat really famous if they take the bull by the horns and take MS on head-on. No longer hinting at calling MS out, but loudly and often calling their bluff enough that it's as obvious to everyone else that MS has nothing to back their bluff as it is too us.
Yes due to the extremely weak and broad patents that have been handed out despite prior art, MS probably has some weak ones. But MS and the whole FOSS community know that they can be worked around if they exist. Since we all know MS is using this as a technique only to create fear, Redhat could be the true American hero by calling them out so publicly and often, that they have to give up. Now's their chance.
Ok there are two facets to this. Yes, the project should do everything it can to make it easy to add good information. The other side is that if someone is so non-confident in their information that they don't back it up or explain why it is an improvement then one shouldn't be surprised that it doesn't stay. Seriously it is so easy to get information into articles. If you're challenged, back it up with a good source. I have contributed to articles that have retained my exact wording for 3-4 years even after significant expansion.
But sometimes it doesn't happen without defending it. The GGP's post is equivalent to getting a few bits off criticism on a research project and abandoning all research forever. Think of the old days of publishing information, you either had to fund it yourself and take on all the rather large risk or find someone willing to do it. The latter often took humdreds of tries with countless rejections. Wikipedia has lowered the bar in terms of difficulty to publishing good information drastically. If that's still not enough for some people, well, there's only so much that can be done.
Yes, the navel gazing is getting seriously out of hand--people place inordinate importance in internal processes. That extends though to worrying about it. The articles are the only really important thing and that has some important implications.
1) By nearly all measures most articles are improving.
2) The content is under a free license
As long as #1 is happening, to an extent who cares about the rest? Well, detractors do, since they have something to whine and complain about. I'm not talking about you ta bu, I'm talking about the people that don't contribute or try poorly and then say, see it doesn't work. But #2 is the real key. Once articles aren't improving some other process can take over the content and improve it. Hopefully a sane stable version system will get implemented soon and that will be the process that keeps the material improving and reduce the impact of vandalism. In any case free content is still working and that's what makes the fact that the trolls come out of the woodwork and get modded up on slashdot so funny.
The other point is that the decrease in new accounts and editing may not mean that good contributions are reducing. For a long time a high proportion of new accounts have been made simply to vandalize. There are several orders of magnitude more accounts than actual contributing accounts. The rest are for garbage and they get blocked and move on to a new account. So 30% less new accounts is meaningless. Also as recent studies (pdf file) have pointed out the anti vandalism bots have significantly impacted vandalism. Perhaps that has deterred a bit of the editing we don't want, which would show up as less editing anyway, making that statistic less important as well.
You're not kidding. This article made me think to go install v 1.6. On a 1.8 Ghz processor running Gnome, Kword for ex opens extremely quickly and opens files quickly as well. This gives me hope that the rest of the codebase is that lean and clean and that it can eventually outdo oo.org. Hopefully it can start to hit critical mass to achieve greater developer mindshare. It's already got oo.org beat in code quality it seems, so hopefully soon in features.
I can certainly say the formula editor is miles ahead of oo.org's in terms of ease of use. I get a font error right away though in starting the formula editor, so I guess I'm off to file a bug report.
If I hadn't commented already, I'd have given you some modpoints. Nice link, especially since there are more courses on the link you gave. The youtube link so far doesn't have many classes, and the lectures are horribly organized with some missing. Hopefully that will improve as the effort moves along. But either way I found a number of courses in the links you gave that I wouldn't mind at all going back to review a bit. In fact the link goes back several years, so you even have the choice of professors in some cases.
Notice how the exam average is 98/240, yet I guarantee that about 90% of the students got an A/A- on their transcript.
This is part of the reason that a science degree from a "top" school means shit these days.
Or that could mean nothing that the average is that low. It's fairly basic testing theory that the harder the test to a point, you get greater dispersion in scores. Then you grade on a curve and can tell who got the A's etc. The difficulty of the test can simply be way more than you really need all the students to know. In other words the average score on the test means nothing without considering how difficult it was. That score could actually mean the students have achieved a good deal of mastery of the material, just not the final 5% or so that doesn't matter so much anyway.
Or the questions on the test if they are very well written, can require a high level of mastery of the course material just to be able to approach them, and instead of testing if you recalled all the facts on page 32 (they basically presuppose that you know all the material), but also ask you to apply and extend it. That's pretty much how my organic chemistry class was. And yes, I can tell you confidently it was much more demanding than those at less prestigious universities. It doesn't mean grading on a curve is the right way to do it, but well, you get the point. The raw number means nothing.
That's true there's always the OCR confidence metric to take into account. What concerns me is that I haven't seen anything that applies random sampling in checking the final accepted answers. What the method description says is if two people agree on a new word it's accepted. Why not scale that number based on the OCR confidence? You mention doing that to reduce the number of people that need to solve it, but why not to increase it? That and/or figure out some procedure to randomly sample accepted answers and send them out again. If a decent percentage of those randomly sampled third tries do not agree with the first two, then you know that 2 isn't the right number for default acceptance. I'm guessing you've thought of this, but I didn't see it anywhere in the description or FAQ.
It is the oldest profession, so of course, it was there the day karma was instituted I'm sure. And people forgot about naked and petrified covered in hot grits. Which actually gives me a good idea on how to win the grand prize. Just have to get Natalie to attend our anniversary party.:)
I read the last link out of curiosity and saw this: "...a specialized computer program which allows you to immediately see, hear, and compare your accent to that of a native Canadian speaker."
I beg of you, is that what we really want, people from developing countries who can perfectly pronounce "hello sir, are you having a chance to get oowt and aboowt of your hoowse today, eh?" Just imagine an accent half Apu and half Saskatchewan and you'll get my drift.
Seriously though, using speech recognition technology to aid in accent reduction is a pretty interesting idea. And as a disclaimer, I'm ethnically Canadian, so doon't get too mad at me.:)
Yes, you've phrased this much better than the way that I hashed it together. Good correction, but is it not true that by a wikipedian's definition a POV backed up by a reliable citation is a fact worth reporting on?
Yes, if the relevant sources support the POV being a prominent one, ie important in some way. But the key thing is that then the view is expressed as what it is and included is who states it, believes it, etc and how prominent it is (to the extent possible). Then prominent criticism, response, etc can be included as appropriate. And for actually admitting on slashdot that someone has pointed out a good correction to something you'd written, I award you a symbolic trophy of your choosing in recognition of the rarity of said event.:)
Not only how long do they last, but how about a full lifecycle analysis of the environmental impact? From production of the solar panels, batteries (if used, if not ignore), etc in a typical installation, to the eventual dispersement of those products either to landfill, recycling, or the land, air, water, etc. From that how do these and other cells compare in general to other power options? I can never find much solid information addressing this question.
It's entirely possible that all the carbon emitted in the production of the solar cells and related equipment is as much or more than the equivalent in other types of power production, especially wind. I realize it's extremely hard to make a comparison between cadmium release and X amount of tons of carbon or other emissions, but how about a stab at it at least? Even just give out the assumptions used and let other people come to other conclusions with other assumptions. At least then the data would be there.
If you've used it often for the things you've mentioned, and not checked further sources each time, then you've probably been taken for a ride at least once. I've been an active contributor for a few years now, and I can tell you that if it is for something important, Wikipedia isn't trustable right now for the most part. And that's fine right now because it's a work in progress, and the only problem is in expecting it to be 100% correct right now. As part of working on articles I see lots of vandalism that goes unseen for too long, inserted misinformation etc, and enough of it to know that if you haven't checked your sources each time, then you've been misled. Now that doesn't mean Wikipedia isn't extremely useful, it can be even now if used right, but more important is that the average article quality seems to be improving. As long as it is improving that's a key piece. But so far we haven't really been able to establish the trustworthiness of the information, something you really need if you need to use the information for something important. That's what the latest innovations are moving towards. With those types of tools, ranging from minimal marking for non vandalized versions all the way up to formally reviewed material, and automatic marking of the trustworthiness of text, the use of Wikipedia articles can get really interesting. If you can keep the free and open editing going while additionally having higher levels of review (and make those easy), then you might really have something.
An important side benefit of the existence of these more formal systems is that it may help make it easier for content experts to get and stay involved in Wikipedia. When they don't have to deal with the low brow vandalism and other forms of idiocy that the radically open system involve, then they may feel more comfortable getting involved. That should be a major goal of the new systems.
Furthermore, I think that you have a fundamental misunderstanding of what wikipedia's purpose is. It has very explicit design goals, using your terms, it attempts to construct articles that have all of the known facts. That it, is ignores "understanding" as you put it, or POV as wiki puts it. If a fact can be attributed to a respectable source then it goes in. Understanding is left as an exercise for the reader.
You had a lot of other good points, but this one is fundamentally wrong. WP does not ignore understanding, or POV or leave them for the reader. What it does, by explicitly chosen policy, is ignore what you and I (or any other editor) think, what our POV is and what our understanding is, unless we publish it in reliable sources. But what a reliable source's understanding of a situation is or their POV on a subject is very fair game (particularly if they are a prominent scholar etc on a subject) and is an essential part of a great article. In fact the presentation of the most important various views on a subject is encouraged. Again, what is against policy for good reason is slanting a POV or making it appear as if a minority view is the dominant one.
Not having to make choices at install time is EXACTLY the reason that ubuntu is good.
I fully agree that not having to make choices at install time is good and it's something that most people that are highly skilled technically cannot understand. But he has a really good point. You can have both if it's done right, and not even confuse the newbie. Make the defaults really easy to choose, in fact, even make it hard to choose anything else, but also make available the advanced options for the people that can install in their sleep and want more options. Perhaps even do it with an initial option that allows an advanced install method. It's not like that's new, lots of distros do it.
I came away from this article with a couple things. One is that's all he found wrong with it? That's good news, since they're not too hard to fix, especially on a vendor installed system. I think he's mostly wrong about the codec issue, as for many codec's Ubuntu is easier than Windows, and after all, how hard is two yes clicks? Another I thought is that the things he pointed out will help improve Ubuntu further. If he has more things that he didn't mention as specific examples, those would be great to hear as that type of good information helps Ubuntu improve.
But the other thing is he's basically acting like Windows is problem free. Well at initial boot a new Windows machine is usually ok. The problem comes within minutes when you have to install AV that bogs the system down and you have to try to maintain the machine as it degrades quickly. Mossberg didn't even appear to try to look for the advantages of Ubuntu, such as the ease of installing new applications without having to separately download and install them as in Windows, and lack of spyware, adware, etc. But it's ok, with those advantages and a free license, Linux and Ubuntu will be fine. We just need to be patient and work on improving it as fast as possible at the same time.
It doesn't matter that it's wrong, some will believe and parrot it. The more energy you spend fighting it, the more people will hear it, and some believe it
This is the best point so far. The point is doesn't matter what he says is true or not. He knows it's not true except in the smallest of senses, and we know it's not true. But because he is at Microsoft and Microsoft is still a respected name especially among businesses (and it is, whether you like it or think it is or not), then some people will still believe it to some extent. He knows that even lying he can influence what people think. The only grain of truth in his statement is that it did require some platform that connected to the internet to make internet search possible. But the main thrust of his statement, that it required MS to be that one is totally false of course. It is of course standard Microsoft public relations: lie and over exaggerate their importance to all technologies. Almost every statement their executives put out in interviews uses it.
Well original research just happens to be the name of the policy, but it covers all unpublished ideas and thought. And what I was saying is that Wikipedia intentionally avoids that type of thing as a necessary evil to maintain improvement in quality. Otherwise you either need a power structure that can say yeah or nay on content or you open floodgates to all the latest crackpot theories and information.You have to spend enough time on the project to reallize there isn't an in between. And again, it's not like there aren't lots of other sources for publishing that other valuable non published information. That's what post-docs are for right?:)
A manual presentation layer. I'm content-driven, personally, a slick presentation does not increase my perception of the value of information.
- Everybody says that, but studies show time and time again that the way information is presented has drastic effects on how much information gets accross and how it is percieved. Next you're going to tell us ads don't affect you.
Right, so it's an automatic (and thus more up-to-date) presentation layer, which carries quantifiable and repeatable bias by virtue of being algorithmic.
- What you're missing here is that google indexes links to information, it does not summarize the actual information as Wikipedia does. Even if the information you wanted was always in a google search, you still then have to collate it and judge sources, etc. Also quality information is not all or perhaps even mostly online right now. The work of summarizing the information is valuable, and if it is already done for you can get you further ahead on the task at hand.
Why should a wiki be "stabilized"? Why is "formality" a virtue when wikipedia was created and gained value from non-conformance to traditional models?
- Because the real goal is information quality. Demonstrable quality in a way useful to the reader/researcher. The non conforming, radically open current system has been shown to be successful in producing content, a smaller portion of it of reasonably high quality. But studies and observation of Wikipedia show that it has extremely high variation in quality. From articles replaced with "YO MAMA SO PHAT..." to widely reviewed articles citing and properly summarizing all the best written material on the subject. Formal peer review can lead to higher information quality and if that reviewed version is available as an option, default or not, can allow the best of both worlds. (like the Linux kernel and most other software) Then there can be both a radically open article that may be more up to date, balanced, etc, and a stable version that is at least guaranteed not to be vandalized. The amount of stabalization could be as little as that or as much as the formally reviewed case, or both. Thus the best of both worlds, content is produced, and high quality content is available, and the review processes can be demonstrated.
Wikipedia has never been interested in unique information. One of the first policies was the one against original research. That certainly doesn't mean there isn't a place for original research, (those are plentiful), nor does it mean Wikipedia isn't valuable. By collating and linking vast amounts of information, Wikipedia does something google can't. It creates the presentation of the information manually. Google can only index content that is already there through an algorithm. And for a long time if not forever, there will be information that is not online. Further, Wikipedia summarizes information like Google will likely never be able to. Even if a Wikipedia article is not all right, it can give you an idea of where to go look and what to look for, which is perhaps it's only truly valuable contribution until there is a way to formally peer review and freeze content so that the reader can see a version that is stabilized.
Two million does sound impressive. Congratulations, Wikipedia. But how does this compare to other encyclopedias? Does anyone have numbers for Britannica or World Book?
Wikipedia does, of course. It turns out some of the largest encyclopedia's written were in Chinese, and that brings in a number of complexities in determining which is larger/est. Of course if you look at all languages of Wikipedia, it's over 8 million now, but that is mostly repeats.
Slashdot Mirror
We go thru this all the time with them, they push prices up to where they get worried we might actually go find an alternative, then bring it down just enough ( but higher then before ) to quiet us down and lose interest in alternatives. Its a cycle that most people are too stupid to see, and thus we are stuck in it.
You're probably right, but lets hope people are a little smarter this time. Maybe this time we have enough momentum and investment in alternative energy research to make the changes. The worst thing for alternative energy projects is that their return on investment calculations are so uncertain because OPEC can just drop oil prices and wipe the project out. What we really need is a higher petroleum tax to offset that effect. Not only will it cover that problem, but it could pay for the alternative energy research and implementation in addition to covering the cost of the externalities (pollution, etc) of the burning of fossil fuels so at least the market is more fair. But of course politicians and people's short sightedness are in the mix, so again, you're probably right. Of course the other possibility is that they really are running out of cheap oil and they aren't bluffing much. We all know it's going to happen sooner or later, it's just the when question.
Seems like a pretty weak test. Or at least not what I was thinking of from the question and lead in reference to Neo.
Experts and extraordinarily talented people in a specific domain do often report the feeling of everything else moving in slow motion. When you're so good at something or have practiced it so many times, you can read the nuances of a situation and react to them so quickly that it's comparable to if everything else had actually slowed down. I've experienced it in a limited way a few times, enough to know what people are talking about when they refer to it.
I can't believe they haven't figured out this is their biggest problem. Not having a specified license after a year means that they are going to be in a license mishmash for years to come. They've already copied Wikipedia articles, so those have to be GFDL. If they go with another license, either they need to get rid of all their GFDL material (a step back) or go forward with multiple licenses and all the incompatibility mess that goes along with that. I for one refuse to contribute until they get the license mess sorted out. And I like the idea of a project where I don't have to put up with all the BS that comes with Wikipedia.
Is trivia not information? - Of course it's information, but that doesn't mean it makes a better article. I submit that the highest quality article is one that is tightly focused on the topic, covers the most important information about the topic, and allocates space in relation to the information's importance to the topic. That way the reader can tell what is important and comes away from an article getting the most important information about the subject. That has the greatest utility to someone coming to use the article for it's information. Yes Wiki is not paper, but that doesn't mean there is space for 40mb of information about a topic all in one article. The less and less important information should be moved to sub articles that are focused on that aspect of a topic. That can go all the way down to the limit of what is verifiable information in credible sources. If that's done properly all the verifiable information is there, it's just organized properly.
The opposite of what I'm getting at is an article with no prioritization of material, random unimportant bits mixed with important ones. That's what trivia bits in articles amount to. Any trivia in an article means more important material has been skipped to give space to the trivia and that results in a lower quality article. Certainly no one is stopping anyone from creating triviapedia where that type of thing would be welcome, but trying to cram trivia in to lower the quality of articles intended to be high quality presentations of information is the problem. Just put it elsewhere, somewhere that's designed for it and there is no problem.
I also submit that what I have proposed here (and is what results when you distill down all of the policy and guideline articles about how to write a great Wikipedia articles) results in self evidently higher quality articles. Again the only conflict comes in trying to use something for something it's not and wondering why there's a conflict.
Exactly. And I'm sure a lot of other people feel the same way. This could make Redhat really famous if they take the bull by the horns and take MS on head-on. No longer hinting at calling MS out, but loudly and often calling their bluff enough that it's as obvious to everyone else that MS has nothing to back their bluff as it is too us.
Yes due to the extremely weak and broad patents that have been handed out despite prior art, MS probably has some weak ones. But MS and the whole FOSS community know that they can be worked around if they exist. Since we all know MS is using this as a technique only to create fear, Redhat could be the true American hero by calling them out so publicly and often, that they have to give up. Now's their chance.
Ok there are two facets to this. Yes, the project should do everything it can to make it easy to add good information. The other side is that if someone is so non-confident in their information that they don't back it up or explain why it is an improvement then one shouldn't be surprised that it doesn't stay. Seriously it is so easy to get information into articles. If you're challenged, back it up with a good source. I have contributed to articles that have retained my exact wording for 3-4 years even after significant expansion.
But sometimes it doesn't happen without defending it. The GGP's post is equivalent to getting a few bits off criticism on a research project and abandoning all research forever. Think of the old days of publishing information, you either had to fund it yourself and take on all the rather large risk or find someone willing to do it. The latter often took humdreds of tries with countless rejections. Wikipedia has lowered the bar in terms of difficulty to publishing good information drastically. If that's still not enough for some people, well, there's only so much that can be done.
Yes, the navel gazing is getting seriously out of hand--people place inordinate importance in internal processes. That extends though to worrying about it. The articles are the only really important thing and that has some important implications.
1) By nearly all measures most articles are improving. 2) The content is under a free license
As long as #1 is happening, to an extent who cares about the rest? Well, detractors do, since they have something to whine and complain about. I'm not talking about you ta bu, I'm talking about the people that don't contribute or try poorly and then say, see it doesn't work. But #2 is the real key. Once articles aren't improving some other process can take over the content and improve it. Hopefully a sane stable version system will get implemented soon and that will be the process that keeps the material improving and reduce the impact of vandalism. In any case free content is still working and that's what makes the fact that the trolls come out of the woodwork and get modded up on slashdot so funny.
The other point is that the decrease in new accounts and editing may not mean that good contributions are reducing. For a long time a high proportion of new accounts have been made simply to vandalize. There are several orders of magnitude more accounts than actual contributing accounts. The rest are for garbage and they get blocked and move on to a new account. So 30% less new accounts is meaningless. Also as recent studies (pdf file) have pointed out the anti vandalism bots have significantly impacted vandalism. Perhaps that has deterred a bit of the editing we don't want, which would show up as less editing anyway, making that statistic less important as well.
You're not kidding. This article made me think to go install v 1.6. On a 1.8 Ghz processor running Gnome, Kword for ex opens extremely quickly and opens files quickly as well. This gives me hope that the rest of the codebase is that lean and clean and that it can eventually outdo oo.org. Hopefully it can start to hit critical mass to achieve greater developer mindshare. It's already got oo.org beat in code quality it seems, so hopefully soon in features.
I can certainly say the formula editor is miles ahead of oo.org's in terms of ease of use. I get a font error right away though in starting the formula editor, so I guess I'm off to file a bug report.
If I hadn't commented already, I'd have given you some modpoints. Nice link, especially since there are more courses on the link you gave. The youtube link so far doesn't have many classes, and the lectures are horribly organized with some missing. Hopefully that will improve as the effort moves along. But either way I found a number of courses in the links you gave that I wouldn't mind at all going back to review a bit. In fact the link goes back several years, so you even have the choice of professors in some cases.
Notice how the exam average is 98/240, yet I guarantee that about 90% of the students got an A/A- on their transcript.
This is part of the reason that a science degree from a "top" school means shit these days.
Or that could mean nothing that the average is that low. It's fairly basic testing theory that the harder the test to a point, you get greater dispersion in scores. Then you grade on a curve and can tell who got the A's etc. The difficulty of the test can simply be way more than you really need all the students to know. In other words the average score on the test means nothing without considering how difficult it was. That score could actually mean the students have achieved a good deal of mastery of the material, just not the final 5% or so that doesn't matter so much anyway.
Or the questions on the test if they are very well written, can require a high level of mastery of the course material just to be able to approach them, and instead of testing if you recalled all the facts on page 32 (they basically presuppose that you know all the material), but also ask you to apply and extend it. That's pretty much how my organic chemistry class was. And yes, I can tell you confidently it was much more demanding than those at less prestigious universities. It doesn't mean grading on a curve is the right way to do it, but well, you get the point. The raw number means nothing.
That's all the same guy (in collaboration with various others). So no wonder you picked up the connection. :)
http://www.cs.cmu.edu/~biglou/research.html
That's true there's always the OCR confidence metric to take into account. What concerns me is that I haven't seen anything that applies random sampling in checking the final accepted answers. What the method description says is if two people agree on a new word it's accepted. Why not scale that number based on the OCR confidence? You mention doing that to reduce the number of people that need to solve it, but why not to increase it? That and/or figure out some procedure to randomly sample accepted answers and send them out again. If a decent percentage of those randomly sampled third tries do not agree with the first two, then you know that 2 isn't the right number for default acceptance. I'm guessing you've thought of this, but I didn't see it anywhere in the description or FAQ.
If you document that on video, I'd hope you could be eligible for prizes. I just never, ever, want to see the video.
It is the oldest profession, so of course, it was there the day karma was instituted I'm sure. And people forgot about naked and petrified covered in hot grits. Which actually gives me a good idea on how to win the grand prize. Just have to get Natalie to attend our anniversary party. :)
I read the last link out of curiosity and saw this: "...a specialized computer program which allows you to immediately see, hear, and compare your accent to that of a native Canadian speaker."
:)
I beg of you, is that what we really want, people from developing countries who can perfectly pronounce "hello sir, are you having a chance to get oowt and aboowt of your hoowse today, eh?" Just imagine an accent half Apu and half Saskatchewan and you'll get my drift.
Seriously though, using speech recognition technology to aid in accent reduction is a pretty interesting idea. And as a disclaimer, I'm ethnically Canadian, so doon't get too mad at me.
Yes, you've phrased this much better than the way that I hashed it together. Good correction, but is it not true that by a wikipedian's definition a POV backed up by a reliable citation is a fact worth reporting on?
:)
Yes, if the relevant sources support the POV being a prominent one, ie important in some way. But the key thing is that then the view is expressed as what it is and included is who states it, believes it, etc and how prominent it is (to the extent possible). Then prominent criticism, response, etc can be included as appropriate. And for actually admitting on slashdot that someone has pointed out a good correction to something you'd written, I award you a symbolic trophy of your choosing in recognition of the rarity of said event.
Not only how long do they last, but how about a full lifecycle analysis of the environmental impact? From production of the solar panels, batteries (if used, if not ignore), etc in a typical installation, to the eventual dispersement of those products either to landfill, recycling, or the land, air, water, etc. From that how do these and other cells compare in general to other power options? I can never find much solid information addressing this question.
It's entirely possible that all the carbon emitted in the production of the solar cells and related equipment is as much or more than the equivalent in other types of power production, especially wind. I realize it's extremely hard to make a comparison between cadmium release and X amount of tons of carbon or other emissions, but how about a stab at it at least? Even just give out the assumptions used and let other people come to other conclusions with other assumptions. At least then the data would be there.
If you've used it often for the things you've mentioned, and not checked further sources each time, then you've probably been taken for a ride at least once. I've been an active contributor for a few years now, and I can tell you that if it is for something important, Wikipedia isn't trustable right now for the most part. And that's fine right now because it's a work in progress, and the only problem is in expecting it to be 100% correct right now. As part of working on articles I see lots of vandalism that goes unseen for too long, inserted misinformation etc, and enough of it to know that if you haven't checked your sources each time, then you've been misled. Now that doesn't mean Wikipedia isn't extremely useful, it can be even now if used right, but more important is that the average article quality seems to be improving. As long as it is improving that's a key piece. But so far we haven't really been able to establish the trustworthiness of the information, something you really need if you need to use the information for something important. That's what the latest innovations are moving towards. With those types of tools, ranging from minimal marking for non vandalized versions all the way up to formally reviewed material, and automatic marking of the trustworthiness of text, the use of Wikipedia articles can get really interesting. If you can keep the free and open editing going while additionally having higher levels of review (and make those easy), then you might really have something.
An important side benefit of the existence of these more formal systems is that it may help make it easier for content experts to get and stay involved in Wikipedia. When they don't have to deal with the low brow vandalism and other forms of idiocy that the radically open system involve, then they may feel more comfortable getting involved. That should be a major goal of the new systems.
Furthermore, I think that you have a fundamental misunderstanding of what wikipedia's purpose is. It has very explicit design goals, using your terms, it attempts to construct articles that have all of the known facts. That it, is ignores "understanding" as you put it, or POV as wiki puts it. If a fact can be attributed to a respectable source then it goes in. Understanding is left as an exercise for the reader.
You had a lot of other good points, but this one is fundamentally wrong. WP does not ignore understanding, or POV or leave them for the reader. What it does, by explicitly chosen policy, is ignore what you and I (or any other editor) think, what our POV is and what our understanding is, unless we publish it in reliable sources. But what a reliable source's understanding of a situation is or their POV on a subject is very fair game (particularly if they are a prominent scholar etc on a subject) and is an essential part of a great article. In fact the presentation of the most important various views on a subject is encouraged. Again, what is against policy for good reason is slanting a POV or making it appear as if a minority view is the dominant one.
Not having to make choices at install time is EXACTLY the reason that ubuntu is good.
I fully agree that not having to make choices at install time is good and it's something that most people that are highly skilled technically cannot understand. But he has a really good point. You can have both if it's done right, and not even confuse the newbie. Make the defaults really easy to choose, in fact, even make it hard to choose anything else, but also make available the advanced options for the people that can install in their sleep and want more options. Perhaps even do it with an initial option that allows an advanced install method. It's not like that's new, lots of distros do it.
I came away from this article with a couple things. One is that's all he found wrong with it? That's good news, since they're not too hard to fix, especially on a vendor installed system. I think he's mostly wrong about the codec issue, as for many codec's Ubuntu is easier than Windows, and after all, how hard is two yes clicks? Another I thought is that the things he pointed out will help improve Ubuntu further. If he has more things that he didn't mention as specific examples, those would be great to hear as that type of good information helps Ubuntu improve.
But the other thing is he's basically acting like Windows is problem free. Well at initial boot a new Windows machine is usually ok. The problem comes within minutes when you have to install AV that bogs the system down and you have to try to maintain the machine as it degrades quickly. Mossberg didn't even appear to try to look for the advantages of Ubuntu, such as the ease of installing new applications without having to separately download and install them as in Windows, and lack of spyware, adware, etc. But it's ok, with those advantages and a free license, Linux and Ubuntu will be fine. We just need to be patient and work on improving it as fast as possible at the same time.
It doesn't matter that it's wrong, some will believe and parrot it. The more energy you spend fighting it, the more people will hear it, and some believe it
This is the best point so far. The point is doesn't matter what he says is true or not. He knows it's not true except in the smallest of senses, and we know it's not true. But because he is at Microsoft and Microsoft is still a respected name especially among businesses (and it is, whether you like it or think it is or not), then some people will still believe it to some extent. He knows that even lying he can influence what people think. The only grain of truth in his statement is that it did require some platform that connected to the internet to make internet search possible. But the main thrust of his statement, that it required MS to be that one is totally false of course. It is of course standard Microsoft public relations: lie and over exaggerate their importance to all technologies. Almost every statement their executives put out in interviews uses it.
Well original research just happens to be the name of the policy, but it covers all unpublished ideas and thought. And what I was saying is that Wikipedia intentionally avoids that type of thing as a necessary evil to maintain improvement in quality. Otherwise you either need a power structure that can say yeah or nay on content or you open floodgates to all the latest crackpot theories and information.You have to spend enough time on the project to reallize there isn't an in between. And again, it's not like there aren't lots of other sources for publishing that other valuable non published information. That's what post-docs are for right? :)
A manual presentation layer. I'm content-driven, personally, a slick presentation does not increase my perception of the value of information.
- Everybody says that, but studies show time and time again that the way information is presented has drastic effects on how much information gets accross and how it is percieved. Next you're going to tell us ads don't affect you.
Right, so it's an automatic (and thus more up-to-date) presentation layer, which carries quantifiable and repeatable bias by virtue of being algorithmic.
- What you're missing here is that google indexes links to information, it does not summarize the actual information as Wikipedia does. Even if the information you wanted was always in a google search, you still then have to collate it and judge sources, etc. Also quality information is not all or perhaps even mostly online right now. The work of summarizing the information is valuable, and if it is already done for you can get you further ahead on the task at hand.
Why should a wiki be "stabilized"? Why is "formality" a virtue when wikipedia was created and gained value from non-conformance to traditional models?
- Because the real goal is information quality. Demonstrable quality in a way useful to the reader/researcher. The non conforming, radically open current system has been shown to be successful in producing content, a smaller portion of it of reasonably high quality. But studies and observation of Wikipedia show that it has extremely high variation in quality. From articles replaced with "YO MAMA SO PHAT..." to widely reviewed articles citing and properly summarizing all the best written material on the subject. Formal peer review can lead to higher information quality and if that reviewed version is available as an option, default or not, can allow the best of both worlds. (like the Linux kernel and most other software) Then there can be both a radically open article that may be more up to date, balanced, etc, and a stable version that is at least guaranteed not to be vandalized. The amount of stabalization could be as little as that or as much as the formally reviewed case, or both. Thus the best of both worlds, content is produced, and high quality content is available, and the review processes can be demonstrated.
Wikipedia has never been interested in unique information. One of the first policies was the one against original research. That certainly doesn't mean there isn't a place for original research, (those are plentiful), nor does it mean Wikipedia isn't valuable. By collating and linking vast amounts of information, Wikipedia does something google can't. It creates the presentation of the information manually. Google can only index content that is already there through an algorithm. And for a long time if not forever, there will be information that is not online. Further, Wikipedia summarizes information like Google will likely never be able to. Even if a Wikipedia article is not all right, it can give you an idea of where to go look and what to look for, which is perhaps it's only truly valuable contribution until there is a way to formally peer review and freeze content so that the reader can see a version that is stabilized.
Two million does sound impressive. Congratulations, Wikipedia. But how does this compare to other encyclopedias? Does anyone have numbers for Britannica or World Book?
Wikipedia does, of course. It turns out some of the largest encyclopedia's written were in Chinese, and that brings in a number of complexities in determining which is larger/est. Of course if you look at all languages of Wikipedia, it's over 8 million now, but that is mostly repeats.