... just like every other standard CAPTCHA system. The key to a good anti-bot protection is to make a question that requires actual intelligence. Most current CAPTCHA systems don't require intelligence, they're just a kind of 'expert system'. There's always some kind of algorithm behind them and this means that any sufficiently motivated programmer can implement this algorithm. The incentive of being able to send massive amounts of spam through something like GMail is very motivating for spammers.
A real Turing test relies on the ability of humans to cope with unknown situations. The amount of 'unknown' in all current CAPTCHA systems is way too small. Any site that is a major potential target for bots should hire a team of dedicated CAPTCHA makers. Their only job is to continuously come up with sufficiently different CAPTCHAs, every few weeks. Each new CAPTCHA is added to a pool or 'battery'. When a new user signs up, a random CAPTCHA is picked from the pool. As soon as one appears to have been cracked, it's withdrawn. Yes, this costs effort, but there's no future in one single CAPTCHA system. It's better to have a whole battery of relatively simple CAPTCHAs than a single complicated one.
No magic required! Just take apart a microwave oven, find a way to focus the microwaves into a narrow bundle, and use a microwave antenna to convert it back to electricity. There you have it, wireless power! Heck, why bundle the waves? Just put an industrial-grade magnetron in the middle of your room and you don't even have to aim the emitter towards the antenna.
Small print: the author of this comment is not responsible for any side effects occurring during this experiment.
Most of these things are an order of magnitude more scandalous than the average Nigerian scam. But it has to be said, it's also an order of magnitude worse to believe in the voodoo of a piece of wood producing GigaHertz waves, than in something which may still be marginally plausible like someone needing to get a large amount of money out of a country.
Your 'trivia' questions are not particularly problematic unless you want to make sure that even 4-year olds or people who can hardly read and write English can post on your forum. Which is something you might not really desire. Even if someone doesn't know the days of the week, or what color a ripe tomato has, looking it up or asking someone by phone or chat is pretty trivial. For a visually impaired person, a captcha is a much higher barrier.
Or if you mean that they would be too easy for a robot to answer, I have yet to see a system that can read and answer any 'trivia' question. If someone builds one, well, that would actually be a useful contribution to computer science.
The main problem with any anti-robot system is that the more standard it becomes, the more rewarding it becomes to crack it. If I write an OCR system that can read CAPTCHAs of a certain kind, all sites using this system become vulnerable. Similarly, if everyone would be using the same set of text questions, spammers will eventually build a database with the answers. But, changing a set of questions is a lot easier and user-friendlier for your visitors than making your CAPTCHAs harder to read. The unicity of your questions is more important than the amount of intelligence required to answer them. For instance, the following list of questions all require the user to simply type an 'a'.
You will have to type a letter 'a' in this field.
Please enter the first letter of the word 'alphabet'.
If my name is Ann, what letter does my name start with?
Type "a" here, without the quotes.
There are 26 letters in the alphabet. What's the first? If you don't know, it's easy to guess from the word itself.
When I say my ABC, what letter do I start with?
What's the last letter in the word 'CAPTCHA'?
I will repeat a certain vowel now: AAAAA. Type it once.
In the following list, one letter is different from the rest, type it. E E A E E E.
Well, this finally makes CAPTCHAs somewhat useful. I won't try to formulate it in some sugar-coated way: I personally hate CAPTCHAs. On some types (especially the ones from Digg), I fail about 50% of them, and that's getting quite annoying after a while. Especially when your code is rejected even if you believe there is no doubt about what you've read in the image.
I believe CAPTCHAs are the wrong solution to the wrong problem. It's a bit exaggerated to call them a "Turing test", because I'm quite sure that OCR systems will be made in the near future that are better than humans in reading CAPTCHAs. A simple text-based question that requires actual intelligence is a much better Turing test, and also a much smaller nuisance for people with impaired vision. Of course, writing a foolproof system that can produce a nearly infinite amount of such questions is a challenging problem by itself.
Isn't it amazing how new technology has enabled people to communicate in a similar way as in the early days of human speech, only over a larger distance and by having to pay for it? True, with SMS/text you can 'talk' with someone hundreds of miles away, but it's slow compared to speech, inefficient (is there any way to do some kind of 'broadcast' with SMS anyway, except for going through your address book or hoping that your friends will pass on the message?), limited, prone to misinterpretation (due to all those abbreviation 'dialects') and very volatile. That's all great for kids and cellphone operators, but I don't see anyone communicating about a project, financial stuff, organization of an event,... over text messages.
That's not the point. The point is that my own playlist doesn't feed me the same irritating commercials every dozen minutes, or spoils the intro & exit of each song with the ramblings of some DJ. I still buy some of the 'new crap' that I hear and like.
Where can I get those tubes of toothpaste and mayonnaise with chip meters that prevent me from squeezing more toothpaste out of the tube when it's supposed to be empty?
I've bought an Epson Stylus 680 long ago, it was one of the first inkjets to have a dreaded chip system mounted on each cartridge, which was supposed to keep track of the ink levels. The printer refuses to work when it thinks the ink is used up. Because genuine Epson cartridges were quite pricey, I bought cheap replacement cartridges. They came with a "chip programmer" device, which allowed to reset the ink meter on the chip. Because the replacement cartridges were delivered without chip, you had to pluck the chip from the original cartridges, and reset it. Of course, I wanted to know how much ink was left in the original cartridges, so put the re-programmed chip back. I printed about 50% of extra pages compared to what I had already printed. This means the chip meter was very pessimistic, wasting 33% of the cartridge's potential. Judging from the article, this seems to have improved. I don't know how they measure the amount of ink used, but I think in the case of my old printer, it just subtracted a very conservative value with each printed page.
Anyhow, I never used that printer much, because every time I turned it on, it probably spat out enough ink to fill 3 pages with solid colors, due to the head cleaning. In the end I tried to stack up print jobs so I could do them all at once. It has now been sitting in a box for so long that every single hole in the print head has probably been clogged.
I wouldn't have any objections against these level meters, if the damn system wouldn't prevent me from printing anything when it thinks it's "empty". This just reeks of ripping off the customers so they have to buy the expensive cartridges more often. Of course, the manufacturers will say that completely emptying the cartridge is bad for the printer heads, so they can sell more cartridges. If it would really be bad, it would be more profitable to make the meter too optimistic, so the heads really get damaged and the user needs to buy a new printer often...
I use a laptop (MacBook Pro C2D 15") as my '24/7' machine (actually more like 15/7). It uses 27W when doing simple stuff like browsing the web, 20W when idle with backlight off, and up to 60W when playing UT2k4. Of course the main reason why it's a laptop, is that it's slightly more portable than my gaming PC.
I expect most energy saving methods currently being used in laptops to be ported to desktops in the near future, because otherwise the power requirements will become unmanageable. If it continues like this, we'll soon have to hook up our PC's to three-phase sockets...
I have made a simulation of my university campus in the game 'Marathon'. You can run around in it and shoot each other with rocket launchers (look for "C200 map marathon" in Google). There's a BB gun in my room, scissors, knives, a soldering iron, and a toothpick. Fear me!
How can a tiny device in which they fitted some cheap optics because there was some room left, take better pictures than a dedicated device based on a proven design with high-quality optics and electronics? In fact the article is not to blame because it doesn't aspire to be a professional camera review and sticks mostly to presenting the facts. The slashdot summary on the other hand reeks of either advertising or the opinion of someone who doesn't really know what he's talking about.
The point is that a good digital camera produces an accurate, high-range representation of the scene, which contains all the information required to turn it into a good final photo. This doesn't necessarily mean that the photo will 'look good' on-screen immediately, it's the job of the photographer to do some postprocessing to make the photo look more 'vibrant', to use our word of the day. The point is that you can easily go from a 'raw' photo to a 'vibrant' one, but not the other way around. Making colours look more vibrant mostly boils down to increasing the contrast and saturation, which means throwing away information. This is exactly what these built-in cameras do: postprocessing is built-in and probably also involves some tricks to compensate for the less than ideal optics and sensor. If they would have provided the full resolution images and done some more tests like outdoor images with white objects in bright sunlight, you'd probably see artefacts like blurring and purple fringing in the corners of the cellphone images. It would also have been interesting to see an attempt at a night shot of a city skyline.
It's also not just the tiny lens that's problematic in a cellphone camera, the sensor also is tiny which means tiny pixels that are prone to noise. Again, seeing the full-resolution images would be interesting to evaluate this. To keep the story short, if you use the same technology as in these phones but upscaled to the size and design of a classic camera, the result will always be better.
Nevertheless, if your only goal is to take pictures on-the-go that look good on average without having to run them through the GIMP or Photoshop, and to post them on Flickr in low-resolution format, then the tiny built-in camera in a device you're always carrying anyway is perfectly adequate. Just don't try to make an awe-inspiring poster from these images or submit them to National Geographic.
Well, with less exhaust gases in the air, there will be less smog. But that's still no excuse for lighting up whole highways for that single car that passes there every 2 minutes.
Yes, it learns that in the future, the whole planet will be littered with them. Which makes sense, if this air car would be successful.
Following the same line of thought, there will also be a need for massive amounts of explosive barrels and crates with medkits.
On a more serious note, this look pretty useful. I've been able to get quite good results on many images with a combination of blurring, denoising, unsharp masking and other algorithms in Photoshop and the GIMP, but nothing beats a proper anisotropic diffusion. And in the versions of the aforementioned software I've used, there's no such thing available (maybe in newer versions, so don't shoot me).
Too bad the inpainting part appears unfinished (at least the documentation says "to do"). Inpainting is the reconstruction of missing parts of the image from surrounding parts. I've managed to do successful inpainting by using a simple diffusion kernel (GIMP and PS allow to specify a custom convolution kernel), but this takes ages because you need to repeat the transformation until the image stabilizes, and you can only use dumb diffusion kernels like a square with a zero in the middle.
... another bot that will eat away the paid bandwidth of my site. Many people have a limited upload quota for their site. Of course they won't put dozens of media files online, but suppose this bot crawls at a quite high rate, a few audio files can quickly gobble up a lot of this quota. Most likely it won't obey robots.txt, so I hope it can be blocked by other means.
This probably involves watermarks that are hidden in the masked part of the spectrum, i.e. in the same way MP3 and similar codecs work. You can't easily remove those without distorting the audio considerably, unless you would know exactly what kind of watermark it is and how to remove it. Of course you can just 'blur' the entire audio clip, but people aren't used to listening to "cassette-tape-that-has-been-lying-in-the-sun-for- too-long" kind of audio anymore.
Of course.
Here's a spectrogram of the first song I ever bought on iTunes: http://homes.esat.kuleuven.be/~athomas/images/am-i tunes.png
A few weeks after I discovered I had the same song on a compilation CD, so I ripped it. This is the spectrogram from the CD-rip: http://homes.esat.kuleuven.be/~athomas/images/am-c d.png
It are not exactly the same parts, plus the spectrograph software I used doesn't always run at the same speed, that's why it looks different. However, you can see that this spectrogram goes all the way up to 20kHz, while the iTunes one appears brickwall-filtered at about 15.5kHz.
It's not so surprising to see this brickwall filter, because any decent AAC or MP3 encoder will by default brickwall filter any audio you try to encode at low bitrates like 128kbps. This is because lowpass filtering is less bad than trying to keep all frequency content in an insufficient bitrate, which will introduce irritating distortion. So it's a compromise, and for many people this quality is sufficient. However, I would expect to get better quality for $1 or 1 Euro per song. These prices are near the price of an actual CD, but the quality isn't. Allofmp3 is much better in this aspect, because you pay per MByte that you download, and the bitrate can be chosen. Until Apple implements a similar scheme, I won't buy anything in the iTMS unless I really, really can't find it elsewhere.
The low bitrate isn't the only thing that keeps me from buying songs from iTunes (unless I really can't find them elsewhere). If you make a spectrogram from songs from the iTunes store, you'll notice that they have no frequency content above 15.5kHz. That's slightly worse than FM broadcast quality. They probably lowpass filter the songs to keep the quality decent within the 128kbps bitrate, or maybe to keep some RIAA bosses happy too.
"The question is, how far down this road do we go before there's any meaningful action to reduce emissions, what does the evidence have to be?" he said. "And unfortunately as human beings -- it doesn't matter really what it is -- we only deal well with crises."
This is so true. The required evidence is a total disaster, as usual. Most people won't do anything about a problem until it really obviously attacks their cozy luxurious way of life or almost kills them. It's useless to explain that a little more effort now can prevent a whole lot of trouble in the future.
Slashdot Mirror
So to recapitulate, stuff gets bigger -> stuff it contains gets more inefficient. Net result is near-zero. Isn't technological advance wonderful?
>>Rapid Onset, Vital Passage and Sudden Thrust. The titles may conjure images of blitzkrieg,
>>
> Sounds more like pr0n.
Or titles for upcoming Jean-Claude Van Damme or Steven Seagal movies.
... just like every other standard CAPTCHA system. The key to a good anti-bot protection is to make a question that requires actual intelligence. Most current CAPTCHA systems don't require intelligence, they're just a kind of 'expert system'. There's always some kind of algorithm behind them and this means that any sufficiently motivated programmer can implement this algorithm. The incentive of being able to send massive amounts of spam through something like GMail is very motivating for spammers.
A real Turing test relies on the ability of humans to cope with unknown situations. The amount of 'unknown' in all current CAPTCHA systems is way too small. Any site that is a major potential target for bots should hire a team of dedicated CAPTCHA makers. Their only job is to continuously come up with sufficiently different CAPTCHAs, every few weeks. Each new CAPTCHA is added to a pool or 'battery'. When a new user signs up, a random CAPTCHA is picked from the pool. As soon as one appears to have been cracked, it's withdrawn. Yes, this costs effort, but there's no future in one single CAPTCHA system. It's better to have a whole battery of relatively simple CAPTCHAs than a single complicated one.
No magic required! Just take apart a microwave oven, find a way to focus the microwaves into a narrow bundle, and use a microwave antenna to convert it back to electricity. There you have it, wireless power! Heck, why bundle the waves? Just put an industrial-grade magnetron in the middle of your room and you don't even have to aim the emitter towards the antenna.
Small print: the author of this comment is not responsible for any side effects occurring during this experiment.
Most of these things are an order of magnitude more scandalous than the average Nigerian scam. But it has to be said, it's also an order of magnitude worse to believe in the voodoo of a piece of wood producing GigaHertz waves, than in something which may still be marginally plausible like someone needing to get a large amount of money out of a country.
Of course I don't mind. It's not like there's a license agreement attached to each of my posts :)
Or if you mean that they would be too easy for a robot to answer, I have yet to see a system that can read and answer any 'trivia' question. If someone builds one, well, that would actually be a useful contribution to computer science.
The main problem with any anti-robot system is that the more standard it becomes, the more rewarding it becomes to crack it. If I write an OCR system that can read CAPTCHAs of a certain kind, all sites using this system become vulnerable. Similarly, if everyone would be using the same set of text questions, spammers will eventually build a database with the answers. But, changing a set of questions is a lot easier and user-friendlier for your visitors than making your CAPTCHAs harder to read. The unicity of your questions is more important than the amount of intelligence required to answer them. For instance, the following list of questions all require the user to simply type an 'a'.
Well, this finally makes CAPTCHAs somewhat useful. I won't try to formulate it in some sugar-coated way: I personally hate CAPTCHAs. On some types (especially the ones from Digg), I fail about 50% of them, and that's getting quite annoying after a while. Especially when your code is rejected even if you believe there is no doubt about what you've read in the image.
I believe CAPTCHAs are the wrong solution to the wrong problem. It's a bit exaggerated to call them a "Turing test", because I'm quite sure that OCR systems will be made in the near future that are better than humans in reading CAPTCHAs. A simple text-based question that requires actual intelligence is a much better Turing test, and also a much smaller nuisance for people with impaired vision. Of course, writing a foolproof system that can produce a nearly infinite amount of such questions is a challenging problem by itself.
Isn't it amazing how new technology has enabled people to communicate in a similar way as in the early days of human speech, only over a larger distance and by having to pay for it? True, with SMS/text you can 'talk' with someone hundreds of miles away, but it's slow compared to speech, inefficient (is there any way to do some kind of 'broadcast' with SMS anyway, except for going through your address book or hoping that your friends will pass on the message?), limited, prone to misinterpretation (due to all those abbreviation 'dialects') and very volatile. That's all great for kids and cellphone operators, but I don't see anyone communicating about a project, financial stuff, organization of an event, ... over text messages.
That's not the point. The point is that my own playlist doesn't feed me the same irritating commercials every dozen minutes, or spoils the intro & exit of each song with the ramblings of some DJ. I still buy some of the 'new crap' that I hear and like.
Where can I get those tubes of toothpaste and mayonnaise with chip meters that prevent me from squeezing more toothpaste out of the tube when it's supposed to be empty?
I've bought an Epson Stylus 680 long ago, it was one of the first inkjets to have a dreaded chip system mounted on each cartridge, which was supposed to keep track of the ink levels. The printer refuses to work when it thinks the ink is used up. Because genuine Epson cartridges were quite pricey, I bought cheap replacement cartridges. They came with a "chip programmer" device, which allowed to reset the ink meter on the chip. Because the replacement cartridges were delivered without chip, you had to pluck the chip from the original cartridges, and reset it. Of course, I wanted to know how much ink was left in the original cartridges, so put the re-programmed chip back. I printed about 50% of extra pages compared to what I had already printed. This means the chip meter was very pessimistic, wasting 33% of the cartridge's potential. Judging from the article, this seems to have improved. I don't know how they measure the amount of ink used, but I think in the case of my old printer, it just subtracted a very conservative value with each printed page.
Anyhow, I never used that printer much, because every time I turned it on, it probably spat out enough ink to fill 3 pages with solid colors, due to the head cleaning. In the end I tried to stack up print jobs so I could do them all at once. It has now been sitting in a box for so long that every single hole in the print head has probably been clogged.
I wouldn't have any objections against these level meters, if the damn system wouldn't prevent me from printing anything when it thinks it's "empty". This just reeks of ripping off the customers so they have to buy the expensive cartridges more often. Of course, the manufacturers will say that completely emptying the cartridge is bad for the printer heads, so they can sell more cartridges. If it would really be bad, it would be more profitable to make the meter too optimistic, so the heads really get damaged and the user needs to buy a new printer often...
I use a laptop (MacBook Pro C2D 15") as my '24/7' machine (actually more like 15/7). It uses 27W when doing simple stuff like browsing the web, 20W when idle with backlight off, and up to 60W when playing UT2k4. Of course the main reason why it's a laptop, is that it's slightly more portable than my gaming PC.
I expect most energy saving methods currently being used in laptops to be ported to desktops in the near future, because otherwise the power requirements will become unmanageable. If it continues like this, we'll soon have to hook up our PC's to three-phase sockets...
I have made a simulation of my university campus in the game 'Marathon'. You can run around in it and shoot each other with rocket launchers (look for "C200 map marathon" in Google). There's a BB gun in my room, scissors, knives, a soldering iron, and a toothpick. Fear me!
How can a tiny device in which they fitted some cheap optics because there was some room left, take better pictures than a dedicated device based on a proven design with high-quality optics and electronics? In fact the article is not to blame because it doesn't aspire to be a professional camera review and sticks mostly to presenting the facts. The slashdot summary on the other hand reeks of either advertising or the opinion of someone who doesn't really know what he's talking about.
The point is that a good digital camera produces an accurate, high-range representation of the scene, which contains all the information required to turn it into a good final photo. This doesn't necessarily mean that the photo will 'look good' on-screen immediately, it's the job of the photographer to do some postprocessing to make the photo look more 'vibrant', to use our word of the day. The point is that you can easily go from a 'raw' photo to a 'vibrant' one, but not the other way around. Making colours look more vibrant mostly boils down to increasing the contrast and saturation, which means throwing away information. This is exactly what these built-in cameras do: postprocessing is built-in and probably also involves some tricks to compensate for the less than ideal optics and sensor. If they would have provided the full resolution images and done some more tests like outdoor images with white objects in bright sunlight, you'd probably see artefacts like blurring and purple fringing in the corners of the cellphone images. It would also have been interesting to see an attempt at a night shot of a city skyline.
It's also not just the tiny lens that's problematic in a cellphone camera, the sensor also is tiny which means tiny pixels that are prone to noise. Again, seeing the full-resolution images would be interesting to evaluate this. To keep the story short, if you use the same technology as in these phones but upscaled to the size and design of a classic camera, the result will always be better.
Nevertheless, if your only goal is to take pictures on-the-go that look good on average without having to run them through the GIMP or Photoshop, and to post them on Flickr in low-resolution format, then the tiny built-in camera in a device you're always carrying anyway is perfectly adequate. Just don't try to make an awe-inspiring poster from these images or submit them to National Geographic.
Well, with less exhaust gases in the air, there will be less smog. But that's still no excuse for lighting up whole highways for that single car that passes there every 2 minutes.
Yes, it learns that in the future, the whole planet will be littered with them. Which makes sense, if this air car would be successful.
Following the same line of thought, there will also be a need for massive amounts of explosive barrels and crates with medkits.
Their examples even include an image of Boba Fett skiing!t ion/img/res_ski.png
http://www.greyc.ensicaen.fr/~dtschump/greycstora
On a more serious note, this look pretty useful. I've been able to get quite good results on many images with a combination of blurring, denoising, unsharp masking and other algorithms in Photoshop and the GIMP, but nothing beats a proper anisotropic diffusion. And in the versions of the aforementioned software I've used, there's no such thing available (maybe in newer versions, so don't shoot me).
Too bad the inpainting part appears unfinished (at least the documentation says "to do"). Inpainting is the reconstruction of missing parts of the image from surrounding parts. I've managed to do successful inpainting by using a simple diffusion kernel (GIMP and PS allow to specify a custom convolution kernel), but this takes ages because you need to repeat the transformation until the image stabilizes, and you can only use dumb diffusion kernels like a square with a zero in the middle.
... another bot that will eat away the paid bandwidth of my site. Many people have a limited upload quota for their site. Of course they won't put dozens of media files online, but suppose this bot crawls at a quite high rate, a few audio files can quickly gobble up a lot of this quota. Most likely it won't obey robots.txt, so I hope it can be blocked by other means.
This probably involves watermarks that are hidden in the masked part of the spectrum, i.e. in the same way MP3 and similar codecs work. You can't easily remove those without distorting the audio considerably, unless you would know exactly what kind of watermark it is and how to remove it. Of course you can just 'blur' the entire audio clip, but people aren't used to listening to "cassette-tape-that-has-been-lying-in-the-sun-for- too-long" kind of audio anymore.
Of course.i tunes.png c d.png
i e-itunes.png i e-cd.png b s-itunes.png
Here's a spectrogram of the first song I ever bought on iTunes:
http://homes.esat.kuleuven.be/~athomas/images/am-
A few weeks after I discovered I had the same song on a compilation CD, so I ripped it. This is the spectrogram from the CD-rip:
http://homes.esat.kuleuven.be/~athomas/images/am-
It are not exactly the same parts, plus the spectrograph software I used doesn't always run at the same speed, that's why it looks different. However, you can see that this spectrogram goes all the way up to 20kHz, while the iTunes one appears brickwall-filtered at about 15.5kHz.
Because this may not be so convincing, I wasted an Euro on buying a song I already had. I didn't only choose this one because of its nice title, but mostly because of the high-frequency content. This is the iTunes spectrogram:
http://homes.esat.kuleuven.be/~athomas/images/bow
As you can see, exactly the same brickwall filtered characteristic. Here's my own rip from exactly the same CD (The Singles Collection):
http://homes.esat.kuleuven.be/~athomas/images/bow
And to top it off, another spectrogram from an iTunes store song. I don't have a CD-version of this, but the appearance of again the same brickwall at 15.5kHz says enough.
http://homes.esat.kuleuven.be/~athomas/images/sca
It's not so surprising to see this brickwall filter, because any decent AAC or MP3 encoder will by default brickwall filter any audio you try to encode at low bitrates like 128kbps. This is because lowpass filtering is less bad than trying to keep all frequency content in an insufficient bitrate, which will introduce irritating distortion. So it's a compromise, and for many people this quality is sufficient. However, I would expect to get better quality for $1 or 1 Euro per song. These prices are near the price of an actual CD, but the quality isn't. Allofmp3 is much better in this aspect, because you pay per MByte that you download, and the bitrate can be chosen. Until Apple implements a similar scheme, I won't buy anything in the iTMS unless I really, really can't find it elsewhere.
The low bitrate isn't the only thing that keeps me from buying songs from iTunes (unless I really can't find them elsewhere). If you make a spectrogram from songs from the iTunes store, you'll notice that they have no frequency content above 15.5kHz. That's slightly worse than FM broadcast quality. They probably lowpass filter the songs to keep the quality decent within the 128kbps bitrate, or maybe to keep some RIAA bosses happy too.