I personally hated having to save EXE attachments, find them, and launch them. Making things harder to do is not a good way to improve security, it only reduces the number of people who can use computers. Autoplay is a good idea. It's no different than RPMs. On a home machine people will type rpm -i without checking scripts or caring where files are placed...
"n-word".com is more interesting than nigger.com. It makes you think. If rob had posted nigger.com I would have immediately dismissed the story (from lack of interest).
After seeing their press release last night, I wrote in with some questions. However, now that they are/.ed, I'm a lot less likely to get a reply so anyone here want to comment on this?
> First, let me say I've been using google for a while now and I really like the results it produce. I noticed you changed your pages today and included a press release which contain information about patents you are seeking :
"Google's PageRankTM technology performs an objective measurement of the importance of web pages that is calculated by solving an equation of 500 million variables and more than 2 billion terms. Google does not determine results by counting links. Instead, Google's PageRank uses the vast link structure of the web as an organizational tool. In essence, Google interprets a link from Page A to Page B as a vote by Page A for Page B. "
Coming from a computer graphics background, this technique sounds similar to integrating radiosity form factors. Here, the surface of a world is covered with "patches." Each patch reflects/emits light and/or heat. Because each patch can possibly reflect energy to every other patch (directly or indirectly) in the solution it becomes nearly the same problem as you have solved.
Instead of using light energy, you substitute "human interest." Obviously you have to have some energy in the system initially before it can propagate. Do you give each site a constant energy and or do use popular hand-picked directories like yahoo?
In the CG world, progressive refinement methods can be done in stages with a time complexity of O(n) that quickly approach their photo-realistic limit in a few iterations.. There are many papers in the ACM SigGraph proceedings over the years regarding this topic, but I'm sure you are well versed in the area of iterative solutions to large linear equation systems. I'm am curious what methods you use? What is 1 unit of "human interest?" I assume you do not attempt to find a perfect solution, what error tolarance do you set? How long does a run take? etc, etc.:) Do you have a paper out somewhere?
------ I am starting to use a similar technique for MP3 "group filtering," somewhat akin to how/. moderates comments. I'm planning a voting system that will be built into my web crawler and also interface with winamp and other popular players, so you don't have to do anything special. Each person is connected in a graph to other people who have voted for the same music.
Music is identified by a music "finger print." The finger prints are calculated by averaging frequency components over the first several non-empty frames of an MP3. This means music can be accurately identified even when encoded with different encoders and at different sample rates (regardless of filename). Time shifting effects introduced by different starting positions become less troublesome because of the averaging.
The result is that you have a personal web crawler that can identify a song after a small portion has been downloaded. If the song is past a threshold of tolerance it won't be downloaded and you can optimize for your taste, bandwidth, and available disk space.
(one reason to post ideas on/. is that it serves as a public record to those who would try to patent it)
I can see it now. Hidden web cams at popular make out spots with ultraviolet flashes. I would like to see one of these on a hand-cam, or in an infered version.
Moderators: You don't moderate something up or down based on weither you agree with it or not. How did this get to be a 5? It contains no information, it's not informative, and it's bairly insightful.
Software isn't really the big cost. If you compare $100 for windows with the salary of a tech support person it almost doesn't matter.
There needs to be objective studies on the cost of tech support for Linux versus Windows. These studies could probably get government funding if anyone is intersted in conducting them.
Is this just speculation, or do you have first hand experience? In my opinion one-time pads are *less* secure than public-key or symmetrical block ciphers. This is because the one-time pad must be protected and carried around, while a single key can be easily memorized. Yes, you can use a block cipher to protect your one time pad, but if you trust block ciphers for that, why not trust it all the way?
Most compromises of secret information involve people - not fancy algorithms, and one-time pads are a real people problem. They are easily lost, destroyed, stolen, or exhausted.
That's interesting, I didn't know there was a 26 character limit.
You can get around this limit with subdomains.
the.artist.formerly.known.asprince.com
(asprince.com is available right now! - and I doubt prince has trademarked it):)
This is one of the strategies to find buffer overflows in servers. Create a really long subdomain name and surf around the net and you'd be surpised at how many cgis you crash. Is there a limit to how long a subdomain name can be or how many deep you can go?
Heh. I sleep next to a server bank. At first the fans and the hard drive activity made it hard to sleep (considering none of the machines are in cases), but now I find it oddly comforting.
I don't know the answer. But, I'm pretty sure that simply reflecting RF energy would not cause a laser-like effect. It would probably be more like a flash-light. The image of the home-built system seems to be shaped this way for the effect of some directionism and focusing.
They also chose to use a non-standard CD density, in part I assume, to thwart piracy. While it's probably a very bad idea to release a system where PCs can copy CDs for 50 cents a pop, I wonder how economies of scale will effect this decision in the long wrong - in terms of price, availability, and error control in the manufacturing processes.
Probably best it keep it to just an ID on there. That way an online database can always have the latest info (new address, hair color, etc). If you have any info on the card you need to also have a signature by the state to verify the information is correct. A good signature using elliptical curves GF(255) would be around 512bits. Adding another 64 bytes.
I live in California now! But, yes they make your digitize your thumbprint here. In Texas I refused to have this done and got away with it somehow. But here, I decided it isn't worth the hassle of fighting. I did consider ways of giving them bad data, like putting glue on my thumb.
It's not that it's a privacy issue, and it's not like I'm a criminal. But... if I ever wanted to become a crimincal, I'd rather not have my fingerprints on file. hehe. It's not an option I'm considering, but I'd still like to have that option.:)
The smallest I've seen a wavelet/fractal compressed fingerprint is 1k, but it's questionable how accurately you can match that against another. It's pretty blurry at 1k. However, I imagine you could make a special fingerprint compression algorithm that stores only key data points in less than 1k. I don't know much about fingerprint recognition though.
I predict fake ids for under 21 people will become more difficult to fake. When bars have access to this database, they can just swipe your cards magnetic strip and get an instant picture and birthdate on screen.
How much information can you reliably store on a magnetic strip on a card anyway? Is it possible to encode an entire picture (10-20k?)
Why waste 5 hours a day putting on makeup when you can paint dots on your face and then let a computer render your face for you. Then you can have faces that are physically impossible otherwise. Some people might complain you can't capture as much human emotion from a CG render, but some of those faces on Star* have so much latex they can hardly smile. There are some problems with occlusion and lighting, but it shouldn't be impossible to have one of the lead characters designed this way.
Rembember when Odo first came out and did the morphing bit? It caught your attention.
Anonymous cowards hoping to protect themselves are very trusting or very uninformed. Your login name says nothing about who you are unless you want it to. It's your cookies, IP address, and date/time of posting that are the real indicators. Rob says he doesn't keep logs, but how can one verify this? For someone supported by advertising, the temptation for keeping logs is just to great. Further, if people start saying liable comments, the powers that be might try to force him to keep logs.
To truely post anonymously, use something like http://www.anonymizer.com. If you are extremely paranoid, don't post.:)
The real reason I see for anonymous cowards is for people who don't want to jump through the hoop of login. Most people are lazy more than they are concerned about their privacy. I've never read any articles on NY Times because they make you register. I used slashdot for almost a year before I registered, but before that I made some informative post as AC - most of which I included my name at the bottom of the post.
This article mentions at least one Russian spy sat takes pictures onto film and drops it off by parachute. In today's digital-everything age, this sound really strange. Obviously the sats are pretty old, but it sounds like they are still operational. How much film (and parachutes) can you stick on one of those things? Do they have a missions for reloading them? I've never heard of anything like that. Anyone with more info on this, please post. Thanks!
Their web site says they might show banner ads to help lower the cost.
Pretty soon people will be giving away free fridges if you allow them to track what you eat. Talk about invasion of privacy.
I'm I wierd? Because I don't spend any time looking at my fridge. I don't see why this would be useful, with ot without banner ads. Though it would be nice to hook up with a food delivery service so you can press a button the second you are out of ice cream. Still, I'd prefer to do that on a PC where the everything is nicer (faster, more resolution, a mouse, you can save receipts, etc).
Slashdot Mirror
I personally hated having to save EXE attachments, find them, and launch them. Making things harder to do is not a good way to improve security, it only reduces the number of people who can use computers. Autoplay is a good idea. It's no different than RPMs. On a home machine people will type rpm -i without checking scripts or caring where files are placed...
"n-word".com is more interesting than nigger.com. It makes you think. If rob had posted nigger.com I would have immediately dismissed the story (from lack of interest).
After seeing their press release last night, I wrote in with some questions. However, now that they are /.ed, I'm a lot less likely to get a reply so anyone here want to comment on this?
:) Do you have a paper out somewhere?
/. moderates comments. I'm planning a voting system that will be built into my web crawler and also interface with winamp and other popular players, so you don't have to do anything special. Each person is connected in a graph to other people who have voted for the same music.
/. is that it serves as a public record to those who would try to patent it)
> First, let me say I've been using google for a while now and I really like the results it produce. I noticed you changed your pages today and included a press release which contain information about patents you are seeking :
"Google's PageRankTM technology performs an objective measurement of the importance of web pages that is calculated by solving an equation of 500 million variables and more than 2 billion terms. Google does not determine results by counting links. Instead, Google's PageRank uses the vast link structure of the web as an organizational tool. In essence, Google interprets a link from Page A to Page B as a vote by Page A for Page B. "
Coming from a computer graphics background, this technique sounds similar to
integrating radiosity form factors. Here, the surface of a world is covered with "patches."
Each patch reflects/emits light and/or heat. Because each patch can possibly reflect energy to
every other patch (directly or indirectly) in the solution it becomes nearly the same problem as you have solved.
Instead of using light energy, you substitute "human interest." Obviously you have to have some energy in the system initially before it can propagate. Do you give each site a constant energy and or do use popular hand-picked directories like yahoo?
In the CG world, progressive refinement methods can be done in stages with a time complexity of O(n) that quickly approach their photo-realistic limit in a few iterations.. There are many papers in the ACM SigGraph proceedings over the years regarding this topic, but I'm sure you are well versed in the area of iterative solutions to large linear equation systems. I'm am curious what methods you use? What is 1 unit of "human interest?" I assume you do not attempt to find a perfect solution, what error tolarance do you set? How long does a run take? etc, etc.
------
I am starting to use a similar technique for MP3 "group filtering," somewhat akin to how
Music is identified by a music "finger print." The finger prints are calculated by averaging frequency components over the first several non-empty frames of an MP3. This means music can be accurately identified even when encoded with different encoders and at different sample rates (regardless of filename). Time shifting effects introduced by different starting positions become less troublesome because of the averaging.
The result is that you have a personal web crawler that can identify a song after a small portion has been downloaded. If the song is past a threshold of tolerance it won't be downloaded and you can optimize for your taste, bandwidth, and available disk space.
(one reason to post ideas on
amen brother
I can see it now. Hidden web cams at popular make out spots with ultraviolet flashes. I would like to see one of these on a hand-cam, or in an infered version.
Moderators: You don't moderate something up or down based on weither you agree with it or not. How did this get to be a 5? It contains no information, it's not informative, and it's bairly insightful.
Software isn't really the big cost. If you compare $100 for windows with the salary of a tech support person it almost doesn't matter.
There needs to be objective studies on the cost of tech support for Linux versus Windows. These studies could probably get government funding if anyone is intersted in conducting them.
$50,000 would be cheap if you could continue exporting.
Is this just speculation, or do you have first hand experience? In my opinion one-time pads are *less* secure than public-key or symmetrical block ciphers. This is because the one-time pad must be protected and carried around, while a single key can be easily memorized. Yes, you can use a block cipher to protect your one time pad, but if you trust block ciphers for that, why not trust it all the way?
Most compromises of secret information involve people - not fancy algorithms, and one-time pads are a real people problem. They are easily lost, destroyed, stolen, or exhausted.
That's interesting, I didn't know there was a 26 character limit.
:)
You can get around this limit with subdomains.
the.artist.formerly.known.asprince.com
(asprince.com is available right now! - and I doubt prince has trademarked it)
This is one of the strategies to find buffer overflows in servers. Create a really long subdomain name and surf around the net and you'd be surpised at how many cgis you crash. Is there a limit to how long a subdomain name can be or how many deep you can go?
Heh. I sleep next to a server bank. At first the fans and the hard drive activity made it hard to sleep (considering none of the machines are in cases), but now I find it oddly comforting.
I would shoot anyone suggesting we use a NT TS solution.
I suggest you use a NT TS solution.
*standing back*
(just checking)
You could even make the device look like a cell phone itself, so that ....
If you do, please do not hold it next to your head when in use. You would be emmiting a lot more energy than what is consider safe.
Good question.
I don't know the answer. But, I'm pretty sure that simply reflecting RF energy would not cause a laser-like effect. It would probably be more like a flash-light. The image of the home-built system seems to be shaped this way for the effect of some directionism and focusing.
They also chose to use a non-standard CD density, in part I assume, to thwart piracy. While it's probably a very bad idea to release a system where PCs can copy CDs for 50 cents a pop, I wonder how economies of scale will effect this decision in the long wrong - in terms of price, availability, and error control in the manufacturing processes.
Yeah, sounds exactly like google. Maybe that's where they started.
Probably best it keep it to just an ID on there. That way an online database can always have the latest info (new address, hair color, etc). If you have any info on the card you need to also have a signature by the state to verify the information is correct. A good signature using elliptical curves GF(255) would be around 512bits. Adding another 64 bytes.
I live in California now! But, yes they make your digitize your thumbprint here. In Texas I refused to have this done and got away with it somehow. But here, I decided it isn't worth the hassle of fighting. I did consider ways of giving them bad data, like putting glue on my thumb.
:)
It's not that it's a privacy issue, and it's not like I'm a criminal. But... if I ever wanted to become a crimincal, I'd rather not have my fingerprints on file. hehe. It's not an option I'm considering, but I'd still like to have that option.
The smallest I've seen a wavelet/fractal compressed fingerprint is 1k, but it's questionable how accurately you can match that against another. It's pretty blurry at 1k. However, I imagine you could make a special fingerprint compression algorithm that stores only key data points in less than 1k. I don't know much about fingerprint recognition though.
Plus the RSA patent expires next year, so it's not that big of a deal.
I predict fake ids for under 21 people will become more difficult to fake. When bars have access to this database, they can just swipe your cards magnetic strip and get an instant picture and birthdate on screen.
How much information can you reliably store on a magnetic strip on a card anyway? Is it possible to encode an entire picture (10-20k?)
Why waste 5 hours a day putting on makeup when you can paint dots on your face and then let a computer render your face for you. Then you can have faces that are physically impossible otherwise. Some people might complain you can't capture as much human emotion from a CG render, but some of those faces on Star* have so much latex they can hardly smile. There are some problems with occlusion and lighting, but it shouldn't be impossible to have one of the lead characters designed this way.
Rembember when Odo first came out and did the morphing bit? It caught your attention.
Anonymous cowards hoping to protect themselves are very trusting or very uninformed. Your login name says nothing about who you are unless you want it to. It's your cookies, IP address, and date/time of posting that are the real indicators. Rob says he doesn't keep logs, but how can one verify this? For someone supported by advertising, the temptation for keeping logs is just to great. Further, if people start saying liable comments, the powers that be might try to force him to keep logs.
:)
To truely post anonymously, use something like http://www.anonymizer.com. If you are extremely paranoid, don't post.
The real reason I see for anonymous cowards is for people who don't want to jump through the hoop of login. Most people are lazy more than they are concerned about their privacy. I've never read any articles on NY Times because they make you register. I used slashdot for almost a year before I registered, but before that I made some informative post as AC - most of which I included my name at the bottom of the post.
This article mentions at least one Russian spy sat takes pictures onto film and drops it off by parachute. In today's digital-everything age, this sound really strange. Obviously the sats are pretty old, but it sounds like they are still operational. How much film (and parachutes) can you stick on one of those things? Do they have a missions for reloading them? I've never heard of anything like that. Anyone with more info on this, please post.
Thanks!
Thanks for posting that one.
It almost looks like a CG render. Hard to tell what is real and what is not these days.
Their web site says they might show banner
ads to help lower the cost.
Pretty soon people will be giving away free fridges if you allow them to track what you eat. Talk about invasion of privacy.
I'm I wierd? Because I don't spend any time looking at my fridge. I don't see why this would be useful, with ot without banner ads. Though it would be nice to hook up with a food delivery service so you can press a button the second you are out of ice cream. Still, I'd prefer to do that on a PC where the everything is nicer (faster, more resolution, a mouse, you can save receipts, etc).