Yeah, it's fair to say that it has improved. I use Ubuntu and an old Fedora box, and the improvements in Ubuntu media support are palpable; Fedora's probably gotten better since I bothered to update as well.
But my point is that there is *no* platform that supports every media type out of the box; I think that's pretty fair to say. That's largely an artifact of the fact that open formats haven't spread enough.
Flash is weird; I actually think it's pretty useful, aside from the obnoxious ads. But it's not really open. I would feel the same way about Wikimedia investing in some sort of open source Flash IDE, for instance. While Flash (and Quicktime, for that matter) are good software, I don't think it jibes with Wikimedia's principles to invest in closed formats.
Why should Wikimedia invest in extensions to iTunes to support free formats? Apple doesn't _like_ free formats. So what guarantee is there that such extensions would have a shelf life at all?
Media is a pain in the ass on every platform. Linux users cringe every time they see a Quicktime file, a Flash file, etc, etc, etc.
Given that state of affairs, it doesn't make sense for an organization that supports freeing information to invest in software from a company that's exacerbating the problem in the first place.
I think you're missing the point of what Ndesanjo Macha said in the interview:
"When it comes to producing information, we don't want to be dependent."
There are at least some Swahili speakers who don't want to use English all the time. And on the flip side, there are people who speak English just fine, and want a Wikipedia in their own language (Welsh, for instance).
What matters isn't "efficiency" or "degree of worldwide readability" or any other such metric. What matters is that the Wikipedia project is committed to openness and to helping any language community, no matter how small at the beginning, start a Wikipedia:
'Wales has described Wikipedia as "an effort to create and distribute a multilingual free encyclopedia of the highest possible quality to every single person on the planet in their own language."'
That means that if you think that all Wikipedias should be in English, even if you're well-intentioned in that belief, then it's not the project for you. Don't get me wrong, I'm sure you have the best interest of Africans at heart when you state that you think it would be better if all Africans learned English. But we might as well say "Hey everybody on planet Earth, let's all speak Volapük!" (Although there are 117,966 articles in Esperanto... I digress.)
As the bureaucrat from the Vietnamese wikipedia mentioned in another comment, it's not possible to predict when a language will hit the critical mass that's necessary for a Wikipedia to start gaining momentum. Now, I don't have a clue what this statistics page on the Vietnamese Wikipedia is saying, but the digits tell the story, right?
It's the same situation for African languages. Amharic, for instance, a language with something like 30 million speakers, from one of the poorest countries on the planet (Ethiopia), nonetheless has a Wikipedia that has made significant progress over the last couple of years (this also in spite of the fact that the writing system of Amharic presents significant technical hurdles for a potential contributor to overcome -- keyboard layouts, etc). Now there are 412 articles. Not a lot, but something -- and growing. Slowly, but surely. And it's speeding up.
Africa is actually farther ahead on Wikipedia than North and South America: Quechua and Nahuatl have just a couple hundred articles each. Navajo, the biggest native language in my country, the US, doesn't seem to have many at all.
But there's a front page!
You can't predict what will happen with those Wikipedias, and it makes no sense to simply rule out the possibility of any one of them picking up steam. And besides (I'm not going to be PC) it's not your decision to make, and that's how Wikipedia works.
It may be too technical now, but there is already some evidence that people are starting to grok it. Or at least, starting to understand the concept of compiling various sources into a single page.
My Yahoo is a good example, and Bloglines, while slightly techier, has also been successful.
One could have argued that HTML was "too technical" back in the day.
The thing is, I think these "free place to read" cafes are really a dead end from a business point of view.
People come in and they "curl up" in a whole chair for hours on end. And sometimes they just buy one drink. And they sit there and read.
Talk about anti-social! One time I tried to start up a conversation with this girl who was reading, and she actually got pissed at me! I mean, seriously, aren't cafes supposed to be places where you, you know, talk?
I think what we need to do is implement some sort of "page number per drink" or "chapter per drink" limit. Obviously, some real hardcore literary types will bring in heavy Russian novels in a tiny font, but overall I think this would really cut down on the "free reading" freeloaders.
"But I'm scared of a technology like this being used against people..."
If you want to see technology being used against people, to kill people, actually, in massive numbers, every day, you need look no further than your steering wheel.
That's why this competition matters: first it's a "desert race," but sooner or later these things will be driving us around.
I eagerly await the day I can read Slashdot while my car drives for me.
There was an episode of 60 Minutes about Google that addresses this:
Defining Google.
And that's crucial: as well-fed and casual as they may look, the folks at Google are intense, burn-the-midnight-flourescent workaholics, all trying to come up with Google's "next big thing."
Google engineer Alan Eustace explains, "One of the ideas that we're working on is machine translation. We strongly believe that there's enough data on the Web and in the world right now to allow us to automatically translate from one language to another."
I have a sinking feeling that once Google starts doing this, they'll blow Language Weaver, Systran, and everyone else on the planet out of the water, because they have the most parallel text to train on.
I just bought some refurbished cartridges at Staples for my HP printer, and they work fine. So, wussup with that? I don't see a date on the packaging anywhere...
Totally missing the point. Yes, there are arrogant bloggers who puff themselves up to the rotundity of the Goodyear blimp.
So what?
Blogging is about *scale*. And no, USENET is not a valid analogy. There are more bloggers now than there were ever people on USENET. Format does matter -- the point is that people who don't know HTML from a salad fork can now express their opinions on whatever they want, whenever they want.
Daily Kos is certainly full of himself. But he's right about this: "dozens (hundreds) of people waging open source journalism can sometimes be more effective than understaffed newsrooms filled with overworked reporters trying to meet deadline."
Bloggers are fact checkers on the media, and it only works because it's easy enough for many, many people to participate.
As for the idea that open source is all about collaboration whereas blogging is all about flame wars, well... read LKML some time. 8^)
"They're not getting excited about using the internet anymore," Barry Wellman, a University of Toronto cyberspace researcher, said of the findings. "Remember when cars came out, and people would say, 'Wow, we're going for a ride today!'
And now some underinformed braindumping, feel free to rip me a new one, perhaps some light will shine in through the oriface:
Isn't it the case that distinct file types are pretty easy to distinguish automatically? After all, the "file" command works.
When you try to distinguish character sets, for instance, you build statistical models of the sets you wish to distinguish, and then compare the unknown text to those.
Couldn't the same be done to any file type? When the application writes the file, the "filesystem" (or whatever it should be called) stamps that new file with the model that it matched -- if it looked like an mpeg, it's an mpeg. It it can't figure it out, it tells you, you would have to take the step of telling it yourself.
But wouldn't that be better than specifying filetypes in the millions of tiny drop-down menus like we do now?
Mother nature faces different engineering challenges than humans do. Since forces affect objects with different relative strengths at different scales (e.g., surface tension vs. gravity), and since nature is mostly based around cellular subunits, there are different sets of restrictions on natural and human designs.
I'm stealing these ideas from Steven Vogel's "Cats Paws and Catapults," which I'm halfway through, heh.
But the idea of round trip translation is one of my pet peeves, because it's pointless, and *really* lousy way to evaluate machine translation.
Consider: if you were to hire a human translator, who specialized in, say, translating English into Spanish, would you want proof of their skill in translating Spanish into English?
Those are two different tasks, ask any translator.
Now, considering that everyone acknowledges that human translation is better than machine translation (duh), does it make any sense to criticize machine translation for doing something even a qualified human may not do well?
I bet that if you somehow strongarmed a human translator into doing a round-trip translation (English->Spanish->English) (without the benefit of the original text for the second step, of course), you'd find significant changes in meaning. Not ungrammatical changes, certainly, but semantic changes.
The reason this drives me nuts is that it shows up in the media all the time as though it were a valid way to evaluate MT. And it isn't.
Evaluating MT is hard; if you're interested in serious measures, check out this guy's work:
http://www.research.ibm.com/people/k/kishore/
The first paper, the one about BLEU, is something of a standard metric.
Yes, these are funny.
But the idea of round trip translation is one of my pet peeves, because it's pointless, and *really* lousy way to evaluate machine translation.
Consider: if you were to hire a human translator, who specialized in, say, translating English into Spanish, would you want proof of their skill in translating Spanish into English?
Those are two different tasks, ask any translator.
Now, considering that everyone acknowledges that human translation is better than machine translation (duh), does it make any sense to criticize machine translation for doing something even a qualified human may not do well?
I bet that if you somehow strongarmed a human translator into doing a round-trip translation (English->Spanish->English) (without the benefit of the original text for the second step, of course), you'd find significant changes in meaning. Not ungrammatical changes, certainly, but semantic changes.
The reason this drives me nuts is that it shows up in the media all the time as though it were a valid way to evaluate MT. And it isn't.
Evaluating MT is hard; if you're interested in serious measures, check out this guy's work:
http://www.research.ibm.com/people/k/kishore/
The first paper, the one about BLEU, is something of a standard metric.
I actually have an IBM M keyboard, sitting right here.
But good god, where do you STICK that thing on the end of the cable? It's got some kind of insane 5-pin plug in it that looks like it's intended for high voltage.
Okay, wait, now that I look at it, it's not an M. But it *is* clicky. Are there adapters for such things?
From the report (page
"...to dramatize only slightly, they are comparable to producting a 747 or ocean liner from the mechanical equivalent of a single fertilized egg."
Huh?
What they are trying to get at here? To me, the analogy suggests two wrong-headed ideas: first that nanotech involves a biological element, which it doesn't, or second, that the amount of matter in an egg can be transformed into something the size of a 747.
I mean, I guess you could use the molecules of an egg to make an airplane, but you'd have to start with one hell of an omelette!
Suuuure I Google. I Google a LOT. About NLP, in fact.
Maybe I didn't phrase the question so clearly: it's just my impression that your average opensource hacker isn't interested in NLP. I don't know whether this is because they aren't interseted in it, think it's boring, or what. That's what I was wondering.
And karma? Eh? What's that?
You're absolutely right about the importance of definitions. Your code is funny (and it's not too far from what Eliza does, for instance), but actually, it's kind of surprising how "sensible" random content can seem if it's taken from a large enough database (like the internet.)
Neuro-Linguistic Programming is really different. It's actually something of a sore spot, I'd say, within the Natural Language Processing world, that this other field, which overlaps only under the most generous of definitions, has swiped some of the mindshare of the acronym. To be honest, the NeuroLinguistic Programming thing seems a little iffy to me. But I'll withhold judgement. The bottom line is that its goals are quite different from those of Computational Linguistics.
(Also: I'm glad you think the site is interesting. It is my site, but I didn't want to shamelessly promote it.)
Slashdot Mirror
Yeah, it's fair to say that it has improved. I use Ubuntu and an old Fedora box, and the improvements in Ubuntu media support are palpable; Fedora's probably gotten better since I bothered to update as well.
But my point is that there is *no* platform that supports every media type out of the box; I think that's pretty fair to say. That's largely an artifact of the fact that open formats haven't spread enough.
Flash is weird; I actually think it's pretty useful, aside from the obnoxious ads. But it's not really open. I would feel the same way about Wikimedia investing in some sort of open source Flash IDE, for instance. While Flash (and Quicktime, for that matter) are good software, I don't think it jibes with Wikimedia's principles to invest in closed formats.
Why should Wikimedia invest in extensions to iTunes to support free formats? Apple doesn't _like_ free formats. So what guarantee is there that such extensions would have a shelf life at all?
Media is a pain in the ass on every platform. Linux users cringe every time they see a Quicktime file, a Flash file, etc, etc, etc.
Given that state of affairs, it doesn't make sense for an organization that supports freeing information to invest in software from a company that's exacerbating the problem in the first place.
I think you're missing the point of what Ndesanjo Macha said in the interview:
There are at least some Swahili speakers who don't want to use English all the time. And on the flip side, there are people who speak English just fine, and want a Wikipedia in their own language (Welsh, for instance).
What matters isn't "efficiency" or "degree of worldwide readability" or any other such metric. What matters is that the Wikipedia project is committed to openness and to helping any language community, no matter how small at the beginning, start a Wikipedia:
That means that if you think that all Wikipedias should be in English, even if you're well-intentioned in that belief, then it's not the project for you. Don't get me wrong, I'm sure you have the best interest of Africans at heart when you state that you think it would be better if all Africans learned English. But we might as well say "Hey everybody on planet Earth, let's all speak Volapük!" (Although there are 117,966 articles in Esperanto... I digress.)
As the bureaucrat from the Vietnamese wikipedia mentioned in another comment, it's not possible to predict when a language will hit the critical mass that's necessary for a Wikipedia to start gaining momentum. Now, I don't have a clue what this statistics page on the Vietnamese Wikipedia is saying, but the digits tell the story, right?
It's the same situation for African languages. Amharic, for instance, a language with something like 30 million speakers, from one of the poorest countries on the planet (Ethiopia), nonetheless has a Wikipedia that has made significant progress over the last couple of years (this also in spite of the fact that the writing system of Amharic presents significant technical hurdles for a potential contributor to overcome -- keyboard layouts, etc). Now there are 412 articles. Not a lot, but something -- and growing. Slowly, but surely. And it's speeding up.
Africa is actually farther ahead on Wikipedia than North and South America: Quechua and Nahuatl have just a couple hundred articles each. Navajo, the biggest native language in my country, the US, doesn't seem to have many at all.
But there's a front page!
You can't predict what will happen with those Wikipedias, and it makes no sense to simply rule out the possibility of any one of them picking up steam. And besides (I'm not going to be PC) it's not your decision to make, and that's how Wikipedia works.
No no, that excerpt was just mistitled. It should have been under "The Beaver".
It may be too technical now, but there is already some evidence that people are starting to grok it. Or at least, starting to understand the concept of compiling various sources into a single page.
My Yahoo is a good example, and Bloglines, while slightly techier, has also been successful.
One could have argued that HTML was "too technical" back in the day.
People come in and they "curl up" in a whole chair for hours on end. And sometimes they just buy one drink. And they sit there and read.
Talk about anti-social! One time I tried to start up a conversation with this girl who was reading, and she actually got pissed at me! I mean, seriously, aren't cafes supposed to be places where you, you know, talk?
I think what we need to do is implement some sort of "page number per drink" or "chapter per drink" limit. Obviously, some real hardcore literary types will bring in heavy Russian novels in a tiny font, but overall I think this would really cut down on the "free reading" freeloaders.
Commies.
If you want to see technology being used against people, to kill people, actually, in massive numbers, every day, you need look no further than your steering wheel.
That's why this competition matters: first it's a "desert race," but sooner or later these things will be driving us around.
I eagerly await the day I can read Slashdot while my car drives for me.
The last term of that formula is (1/16)^n, so if n gets to 83,431... I'm still pretty damn impressed. =)
Their service is indeed excellent. I had a tire split and they mailed me a replacement kit free of charge.
They also tried to sell me some more of the little filters over the phone, but whatever -- I just spray those clean once in a while with canned air.
*sniff* I heart my Roomba.
Mine says 23mL. *shrug*
I just bought some refurbished cartridges at Staples for my HP printer, and they work fine. So, wussup with that? I don't see a date on the packaging anywhere...
Totally missing the point. Yes, there are arrogant bloggers who puff themselves up to the rotundity of the Goodyear blimp.
So what?
Blogging is about *scale*. And no, USENET is not a valid analogy. There are more bloggers now than there were ever people on USENET. Format does matter -- the point is that people who don't know HTML from a salad fork can now express their opinions on whatever they want, whenever they want.
Daily Kos is certainly full of himself. But he's right about this: "dozens (hundreds) of people waging open source journalism can sometimes be more effective than understaffed newsrooms filled with overworked reporters trying to meet deadline."
Bloggers are fact checkers on the media, and it only works because it's easy enough for many, many people to participate.
As for the idea that open source is all about collaboration whereas blogging is all about flame wars, well... read LKML some time. 8^)
Isn't it the case that distinct file types are pretty easy to distinguish automatically? After all, the "file" command works.
When you try to distinguish character sets, for instance, you build statistical models of the sets you wish to distinguish, and then compare the unknown text to those.
Couldn't the same be done to any file type? When the application writes the file, the "filesystem" (or whatever it should be called) stamps that new file with the model that it matched -- if it looked like an mpeg, it's an mpeg. It it can't figure it out, it tells you, you would have to take the step of telling it yourself.
But wouldn't that be better than specifying filetypes in the millions of tiny drop-down menus like we do now?
Mother nature faces different engineering challenges than humans do. Since forces affect objects with different relative strengths at different scales (e.g., surface tension vs. gravity), and since nature is mostly based around cellular subunits, there are different sets of restrictions on natural and human designs.
I'm stealing these ideas from Steven Vogel's "Cats Paws and Catapults," which I'm halfway through, heh.
It's an interesting book.
Yes, these are funny.
But the idea of round trip translation is one of my pet peeves, because it's pointless, and *really* lousy way to evaluate machine translation.
Consider: if you were to hire a human translator, who specialized in, say, translating English into Spanish, would you want proof of their skill in translating Spanish into English?
Those are two different tasks, ask any translator.
Now, considering that everyone acknowledges that human translation is better than machine translation (duh), does it make any sense to criticize machine translation for doing something even a qualified human may not do well?
I bet that if you somehow strongarmed a human translator into doing a round-trip translation (English->Spanish->English) (without the benefit of the original text for the second step, of course), you'd find significant changes in meaning. Not ungrammatical changes, certainly, but semantic changes.
The reason this drives me nuts is that it shows up in the media all the time as though it were a valid way to evaluate MT. And it isn't.
Evaluating MT is hard; if you're interested in serious measures, check out this guy's work:
http://www.research.ibm.com/people/k/kishore/
The first paper, the one about BLEU, is something of a standard metric.
Yes, these are funny. But the idea of round trip translation is one of my pet peeves, because it's pointless, and *really* lousy way to evaluate machine translation. Consider: if you were to hire a human translator, who specialized in, say, translating English into Spanish, would you want proof of their skill in translating Spanish into English? Those are two different tasks, ask any translator. Now, considering that everyone acknowledges that human translation is better than machine translation (duh), does it make any sense to criticize machine translation for doing something even a qualified human may not do well? I bet that if you somehow strongarmed a human translator into doing a round-trip translation (English->Spanish->English) (without the benefit of the original text for the second step, of course), you'd find significant changes in meaning. Not ungrammatical changes, certainly, but semantic changes. The reason this drives me nuts is that it shows up in the media all the time as though it were a valid way to evaluate MT. And it isn't. Evaluating MT is hard; if you're interested in serious measures, check out this guy's work: http://www.research.ibm.com/people/k/kishore/ The first paper, the one about BLEU, is something of a standard metric.
I actually have an IBM M keyboard, sitting right here.
But good god, where do you STICK that thing on the end of the cable? It's got some kind of insane 5-pin plug in it that looks like it's intended for high voltage.
Okay, wait, now that I look at it, it's not an M. But it *is* clicky. Are there adapters for such things?
From the report (page "...to dramatize only slightly, they are comparable to producting a 747 or ocean liner from the mechanical equivalent of a single fertilized egg." Huh? What they are trying to get at here? To me, the analogy suggests two wrong-headed ideas: first that nanotech involves a biological element, which it doesn't, or second, that the amount of matter in an egg can be transformed into something the size of a 747. I mean, I guess you could use the molecules of an egg to make an airplane, but you'd have to start with one hell of an omelette!
Suuuure I Google. I Google a LOT. About NLP, in fact. Maybe I didn't phrase the question so clearly: it's just my impression that your average opensource hacker isn't interested in NLP. I don't know whether this is because they aren't interseted in it, think it's boring, or what. That's what I was wondering. And karma? Eh? What's that?
You're absolutely right about the importance of definitions. Your code is funny (and it's not too far from what Eliza does, for instance), but actually, it's kind of surprising how "sensible" random content can seem if it's taken from a large enough database (like the internet.)
Neuro-Linguistic Programming is really different. It's actually something of a sore spot, I'd say, within the Natural Language Processing world, that this other field, which overlaps only under the most generous of definitions, has swiped some of the mindshare of the acronym. To be honest, the NeuroLinguistic Programming thing seems a little iffy to me. But I'll withhold judgement. The bottom line is that its goals are quite different from those of Computational Linguistics. (Also: I'm glad you think the site is interesting. It is my site, but I didn't want to shamelessly promote it.)