Burning a fossil fuel generates heat energy and CO2. Photosynthesis consumes solar energy and CO2 and creates sugars and oxygen. The amount of CO2 that the green slime can turn back into oxygen is limited by the amount of solar energy that the designers can funnel into that smokestack. To eliminate most or all of the CO2, the energy they bring in has to be equal to the energy generated by the fossil fuel burning. At that point, they've come up with some sort of brilliant solar power scheme and can just chuck the incinerator and the green slime in favor of using the solar power directly.
That is to say, the idea is flawed. Even if they get green slime to survive at high temperatures, they bring sunlight into the smoke stacks, and they think of something clever to do with the carbon generated, they still can't bring in anywhere near enough sunlight to photosynthesize away all the CO2.
Back to the drawing board. Burning coal is still evil.
Novel. Expensive. Too cumbersome for Real Users to bother with. Panel displays will stay cheaper and easier for the foreseeable future.
the equivalent of a 19" monitor
The 19" monitor claim is silly -- it indicates how much of the retina the display covers and nothing else. The other facts tell a different story: 800x600, 32 shades of grey, 60Hz. Blah.
input might be harder - i don't know of anything but chording keyboards
Getting users to want chord keyboards should be about as easy as getting them to want retina-projection HMDs. Good luck.
what's the down side to sub-1lb devices?
Battery power is directly proportional to battery mass. A laptop's ruggedness varies with its weight, as well. The 12" screens and full-size keyboards that people insist on aren't weightless. I can't think of any laptop under 3 pounds that has been any kind of commercial success. Mainstream laptops have been roughly the same size and weight for the last 5 years, despite whatever advances in miniaturization.
just because the current state of the art makes really tiny devices uncomfortable doesn't mean it'll always be that way.
I didn't say it would. My point is that the offerings featured in this story -- the Fiva et al -- are just more of the same. (There's not a single novel input or output mechanism in the lot of them.) They have all of the Libretto's strengths and weaknesses, with a trendy new CPU inside.
For all the fawning and drooling in the linked article over how small and light the demonstrated devices are, the world doesn't want a 6", 2lb notebook. The world largely ignored the Toshiba Libretto and the Sony PictureBook, presumably because small keyboards and small screens do not make for happy users.
Miniaturization has happened quite nicely without the Crusoe, and manufacturers have discovered that people don't want laptops much below 10"x7"x1" or below about 3.5 pounds.
I do own and use a Libretto. It's amazingly portable (roughly the size of a VHS tape), and it turns a lot of heads, but I can't say I'm a fan of 60%-size keyboards or 640x480 screens.
The primary benefit of the Crusoe is its power savings, which has little or nothing to do with miniaturization. Why not put Crusoes in real laptops, which people actually use, and realize the power savings there?
Aggregate bandwidth requirements would change in an Akamai-like caching server system,
Akamai reduces the amount of wide-area bandwidth used for retrieving content. It doesn't reduce the wide-area bandwidth or server load consumed by searches/lookups and meta-information. The original post in this thread was talking about how much bandwidth is consumed by meta-information, which Akamai does nothing to alleviate.
I'm working on it, both analytically and in simulation. Gnutella requires O(N) bandwidth at each host, or O(N^2) overall. Napster requires O(1) bandwidth at each host, or O(N) overall. The single point of failure indeed buys better network scaling.
I'd guess that in a P2P network the bandwidth required to carry meta-information would go up O(N^2)
That depends on the design. Not all designs are as bad as Gnutella. A hierarchical index can provide scalability without being entirely centralized. Caching can probabilistically reduce traffic. Don't assume that just because Gnutella is bad, P2P as a whole is bad.
The Napster architecture, while introducing a single point of failure (at least from a legal standpoint)
Napster's architecture has a single point of failure from reliability and scalability standpoints as well. Illegality is not their only weakness.
centralizes meta information allowing O(N) growth of query bandwidth in nodes, and decentralizes data transfer
In that respect, Napster is directly analogous to how search engines on the web work. Napster has the same scalability problems as search engines, too: there are just too many documents for any one central point to store all the meta information.
Yes, there is a better way. I'll publish it once I have a working prototype.
Bob Metcalfe has also mispredicated the universal collapse of the Internet twice, the collapse of the technology stock market once, and the collapse of Linux and open source once. It's all here. Why does anyone still take his predictions at all seriously?
Okay, so he figured out how to put the ALOHA radio network protocol on a wire called "Ethernet." He based a company on this and got wealthy doing so. That doesn't make him a visionary any more than ALTAIR BASIC makes Bill Gates a visionary or VI makes Bill Joy a visionary.
Ignore the visionaries. Write code and prove Bob wrong. Again.
if I were to rip my Big Daddy albums into MP3s, burn them to a CD, and give the CDs to friends so they can experience Big Daddy's music, I'm breaking the law. Yet if I don't break the law, this music will eventually disappear completely.
You're confusing two different acts: transferring a copyrighted audio work to a different medium, and distributing copies to your friends. The former is legal under the Audio Home Recording Act of 1992, and the latter is not. Of course, at any time you could also fully transfer (as a loan, or permanently) your only copy of the copyrighted work to one of your friends, and that's legal, too.
So if you can hold onto your copy, make other copies for personal use (for backup or for use with a different playback device), and lend your copy to your friends, I don't think you can fairly say that the music is in danger of disappearing. You can't even fairly say that your friends cannot appreciate the music.
I agree with you on principle that no-longer-available works can be copied without harm to the copyright holder, but copyright law overrides principle if you get caught.
If you've taken a algorithms course (and passed) you, too, could probably write an Oracle of Bacon.
I believe that this was done on a single computer. Pretty sure it wasn't a cluster of 4,000;)
The Oracle takes up about 10% of the CPU time on a single Sun Ultra 5/300. (I didn't pick the machine. The Oracle also runs on my Linux 2xP2/350 at home.) It takes around 80 MB of memory -- 25 for the actors and movies and the rest for a cache of recent queries. Each query consumes 0.6 seconds of CPU time, or 0.02 seconds if it comes from the cache. 90-95% of queries get served from the cache, so the Oracle should withstand 10+ queries per second, sustained.
The task is trivially parallelizable across big clusters (UVA has a 256-node cluster that would do the trick), but the need for that has never arisen...:)
8.The company assumes they will recover this loss when these thousands of people start paying for their internet service 9.These users don't sign up for the internet service and the company doesn't make their money back
If the company doesn't require the purchase of Internet service, it's their own blunder. In capitalistic markets, stupid companies die, and smart ones succeed. Pitying dumb companies is counterproductive for yourself and for the market as a whole. Selling products at a loss without some assurance of a tied-in gain is stupid, and any company that does it is asking to fail.
Fortunately for Netpliance (and unfortunately for me), iOpeners are only $99 with premium service, a whopping $22/mo. If they allow users to cancel the service immediately, we're back to "It's their own damn fault" territory.
As a general rule of security, anything that occurs on an untrusted platform produces untrustable results. Apply this to online voting, and you can conclude that online voting may never work. If the clients are millions of home computers, set up by people who don't know anything about security, no amount of nifty cryptography will make the vote packet that the computer sends to the vote server trusted.
Again: as long as home computers are an untrusted platform, any information coming out of them is suspect.
How do you detect coerced voting when you don't have poll watchers? The whole idea of the secret free vote goes down the drain.
Coerced votes are not unique to online voting. Absentee ballots have the same potential for coercion, for the same reasons, and it has turned out not to be too much of an issue. See the letters at the end of the last issue of Crypto-Gram here.
In short, absentee ballots account for as much as 35-50% of the vote in some elections, certainly enough to change the outcome of an election, and they continue to be used despite the possibility of coercion.
Of course, online voting has many, many other problems. I just don't think coercion is one of them.
The DGPL (database GPL) suggestion is orthoganal to, or counter to, the GPL -- I can't decide which.
First off, the GPL and LGPL do not prevent you from charging a distribution fee, limiting access, or associating advertising with downloads from your site. You can put advertisements on the same web page from which you distribute or mirror GPL'd code. You can limit access, or even charge for access.
However, the GPL also explicitly allows (and requires you to allow) people who download the code to set up their own sites to mirror the code, with or without access restrictions, payment, or ads. If I had to pay a penny per page to use your database but I could get it for free from Bob's House of GPL'd Databases, where do you think I'd go?
In short, the DGPL does not suggest the same solution, or even try to solve quite the same problem, as the GPL. To charge per-search fees or tie ads to searches would require that you make content (i.e., search results and the database itself) proprietary, which runs very much counter to the GPL. Many schemes for making content proprietary exist, including not putting any explicit copyright on it at all; the GPL is not such a scheme. Calling suggestions like enforced per-search fees or advertising tie-ins the "DGPL" is misleading.
The closest analogue to the GPL in the world of "content" is probably the Open Content license. Let's stick with that.
You said: they have someone write a binary only kernel module uder NDA
The "someone" is Dale Whitfield, and he is actually not associated with Logitech in any way. He got specs from Connectix to write an OS/2 driver for the Quickcam VC, and I've talked him into porting the driver to Linux. (I've also been helping a little bit with the effort.) Logitech has nothing to do with this driver.
You also said: binary only drivers suck
While I agree, I contend that they're better than nothing in this case. (There's no point in trying to convince Dale to release the source; he would if he could. Bug Logitech to drop the NDA.) If you disagree, don't use them.
You said: The Powers That Be will not release the information needed for mpeg (I believe) encoding, at least not for free.
There are free MPEG encoders out there. Cqcam doesn't support movies because I haven't had the time to add support, and no one else has contributed it.
All useful AVI and MOV formats (e.g., Indeo, Cinepak, or Sorensen) are under NDA, so those won't be supported. There are some simple (basically uncompressed) formats for AVI. Those aren't really worth supporting.
If anyone wants to contribute MPEG movie generation for cqcam, I'd be happy to work it in.
You said: So I can't go and buy a NEW camera (or, in my case, lots of them) and ship a bunch of web cams, can i?
The information on my page is still correct. Nothing made after the Quickcam 2 is supported (yet). Sorry. There are still QC2's available refurbished, and they work well with cqcam and other available drivers.
The latest news is that there is a QCVC/QCPro driver under development. It will be binary-only under NDA (unless Logitech changes their minds about that) and should be available RSN. I'm not doing the development, so I can't give a date. Watch the Quickcam Third Party Drivers mailing list or my Quickcam pages for announcements.
Slashdot Mirror
That is to say, the idea is flawed. Even if they get green slime to survive at high temperatures, they bring sunlight into the smoke stacks, and they think of something clever to do with the carbon generated, they still can't bring in anywhere near enough sunlight to photosynthesize away all the CO2.
Back to the drawing board. Burning coal is still evil.
Novel. Expensive. Too cumbersome for Real Users to bother with. Panel displays will stay cheaper and easier for the foreseeable future.
the equivalent of a 19" monitor
The 19" monitor claim is silly -- it indicates how much of the retina the display covers and nothing else. The other facts tell a different story: 800x600, 32 shades of grey, 60Hz. Blah.
input might be harder - i don't know of anything but chording keyboards
Getting users to want chord keyboards should be about as easy as getting them to want retina-projection HMDs. Good luck.
what's the down side to sub-1lb devices?
Battery power is directly proportional to battery mass. A laptop's ruggedness varies with its weight, as well. The 12" screens and full-size keyboards that people insist on aren't weightless. I can't think of any laptop under 3 pounds that has been any kind of commercial success. Mainstream laptops have been roughly the same size and weight for the last 5 years, despite whatever advances in miniaturization.
just because the current state of the art makes really tiny devices uncomfortable doesn't mean it'll always be that way.
I didn't say it would. My point is that the offerings featured in this story -- the Fiva et al -- are just more of the same. (There's not a single novel input or output mechanism in the lot of them.) They have all of the Libretto's strengths and weaknesses, with a trendy new CPU inside.
Miniaturization has happened quite nicely without the Crusoe, and manufacturers have discovered that people don't want laptops much below 10"x7"x1" or below about 3.5 pounds.
I do own and use a Libretto. It's amazingly portable (roughly the size of a VHS tape), and it turns a lot of heads, but I can't say I'm a fan of 60%-size keyboards or 640x480 screens.
The primary benefit of the Crusoe is its power savings, which has little or nothing to do with miniaturization. Why not put Crusoes in real laptops, which people actually use, and realize the power savings there?
Akamai reduces the amount of wide-area bandwidth used for retrieving content. It doesn't reduce the wide-area bandwidth or server load consumed by searches/lookups and meta-information. The original post in this thread was talking about how much bandwidth is consumed by meta-information, which Akamai does nothing to alleviate.
I'm working on it, both analytically and in simulation. Gnutella requires O(N) bandwidth at each host, or O(N^2) overall. Napster requires O(1) bandwidth at each host, or O(N) overall. The single point of failure indeed buys better network scaling.
I'd guess that in a P2P network the bandwidth required to carry meta-information would go up O(N^2)
That depends on the design. Not all designs are as bad as Gnutella. A hierarchical index can provide scalability without being entirely centralized. Caching can probabilistically reduce traffic. Don't assume that just because Gnutella is bad, P2P as a whole is bad.
The Napster architecture, while introducing a single point of failure (at least from a legal standpoint)
Napster's architecture has a single point of failure from reliability and scalability standpoints as well. Illegality is not their only weakness.
centralizes meta information allowing O(N) growth of query bandwidth in nodes, and decentralizes data transfer
In that respect, Napster is directly analogous to how search engines on the web work. Napster has the same scalability problems as search engines, too: there are just too many documents for any one central point to store all the meta information.
Yes, there is a better way. I'll publish it once I have a working prototype.
Solaris doesn't run much better on x86 than it does on MIPS. :)
Okay, so he figured out how to put the ALOHA radio network protocol on a wire called "Ethernet." He based a company on this and got wealthy doing so. That doesn't make him a visionary any more than ALTAIR BASIC makes Bill Gates a visionary or VI makes Bill Joy a visionary.
Ignore the visionaries. Write code and prove Bob wrong. Again.
You're confusing two different acts: transferring a copyrighted audio work to a different medium, and distributing copies to your friends. The former is legal under the Audio Home Recording Act of 1992, and the latter is not. Of course, at any time you could also fully transfer (as a loan, or permanently) your only copy of the copyrighted work to one of your friends, and that's legal, too.
So if you can hold onto your copy, make other copies for personal use (for backup or for use with a different playback device), and lend your copy to your friends, I don't think you can fairly say that the music is in danger of disappearing. You can't even fairly say that your friends cannot appreciate the music.
I agree with you on principle that no-longer-available works can be copied without harm to the copyright holder, but copyright law overrides principle if you get caught.
That's me.
Building a graph is quite straightforward
If you've taken a algorithms course (and passed) you, too, could probably write an Oracle of Bacon.
I believe that this was done on a single computer. Pretty sure it wasn't a cluster of 4,000 ;)
The Oracle takes up about 10% of the CPU time on a single Sun Ultra 5/300. (I didn't pick the machine. The Oracle also runs on my Linux 2xP2/350 at home.) It takes around 80 MB of memory -- 25 for the actors and movies and the rest for a cache of recent queries. Each query consumes 0.6 seconds of CPU time, or 0.02 seconds if it comes from the cache. 90-95% of queries get served from the cache, so the Oracle should withstand 10+ queries per second, sustained.
The task is trivially parallelizable across big clusters (UVA has a 256-node cluster that would do the trick), but the need for that has never arisen... :)
--Patrick
9.These users don't sign up for the internet service and the company doesn't make their money back
If the company doesn't require the purchase of Internet service, it's their own blunder. In capitalistic markets, stupid companies die, and smart ones succeed. Pitying dumb companies is counterproductive for yourself and for the market as a whole. Selling products at a loss without some assurance of a tied-in gain is stupid, and any company that does it is asking to fail.
Fortunately for Netpliance (and unfortunately for me), iOpeners are only $99 with premium service, a whopping $22/mo. If they allow users to cancel the service immediately, we're back to "It's their own damn fault" territory.
Again: as long as home computers are an untrusted platform, any information coming out of them is suspect.
Coerced votes are not unique to online voting. Absentee ballots have the same potential for coercion, for the same reasons, and it has turned out not to be too much of an issue. See the letters at the end of the last issue of Crypto-Gram here.
In short, absentee ballots account for as much as 35-50% of the vote in some elections, certainly enough to change the outcome of an election, and they continue to be used despite the possibility of coercion.
Of course, online voting has many, many other problems. I just don't think coercion is one of them.
The DGPL (database GPL) suggestion is orthoganal to, or counter to, the GPL -- I can't decide which.
First off, the GPL and LGPL do not prevent you from charging a distribution fee, limiting access, or associating advertising with downloads from your site. You can put advertisements on the same web page from which you distribute or mirror GPL'd code. You can limit access, or even charge for access.
However, the GPL also explicitly allows (and requires you to allow) people who download the code to set up their own sites to mirror the code, with or without access restrictions, payment, or ads. If I had to pay a penny per page to use your database but I could get it for free from Bob's House of GPL'd Databases, where do you think I'd go?
In short, the DGPL does not suggest the same solution, or even try to solve quite the same problem, as the GPL. To charge per-search fees or tie ads to searches would require that you make content (i.e., search results and the database itself) proprietary, which runs very much counter to the GPL. Many schemes for making content proprietary exist, including not putting any explicit copyright on it at all; the GPL is not such a scheme. Calling suggestions like enforced per-search fees or advertising tie-ins the "DGPL" is misleading.
The closest analogue to the GPL in the world of "content" is probably the Open Content license. Let's stick with that.
The "someone" is Dale Whitfield, and he is actually not associated with Logitech in any way. He got specs from Connectix to write an OS/2 driver for the Quickcam VC, and I've talked him into porting the driver to Linux. (I've also been helping a little bit with the effort.) Logitech has nothing to do with this driver.
You also said: binary only drivers suck
While I agree, I contend that they're better than nothing in this case. (There's no point in trying to convince Dale to release the source; he would if he could. Bug Logitech to drop the NDA.) If you disagree, don't use them.
--Patrick
There are free MPEG encoders out there. Cqcam doesn't support movies because I haven't had the time to add support, and no one else has contributed it.
All useful AVI and MOV formats (e.g., Indeo, Cinepak, or Sorensen) are under NDA, so those won't be supported. There are some simple (basically uncompressed) formats for AVI. Those aren't really worth supporting.
If anyone wants to contribute MPEG movie generation for cqcam, I'd be happy to work it in.
--Patrick
The information on my page is still correct. Nothing made after the Quickcam 2 is supported (yet). Sorry. There are still QC2's available refurbished, and they work well with cqcam and other available drivers.
The latest news is that there is a QCVC/QCPro driver under development. It will be binary-only under NDA (unless Logitech changes their minds about that) and should be available RSN. I'm not doing the development, so I can't give a date. Watch the Quickcam Third Party Drivers mailing list or my Quickcam pages for announcements.
--Patrick