Technically you're right. But I'm pretty sure that if some idiot chose "me@mydomain.com" as his personal email address, he's already used to getting mountains of spam.
There is no single Condorcet method. There are several Condorcet methods with various pros and cons, but I think Ranked Pairs manages to be the most fair. It's certainly much fairer than IRV.
Calculating the Ranked Pairs winner can be complicated, but voting is easy. Simply rank the candidates in order of preference, and skip those you don't know/care about. Anyone can follow that.
Approval voting has a nice simplicity to it, but it fails to distinguish between strong approval and weak approval, and I think it creates much more temptation for people to game the system by how they vote.
Whatever the system chosen, it would be a huge help in changing politics for the better. Unfortunately, it's not likely to happen in our lifetimes.
[To a] GPL advocate... it is the code that comes first, and the applications built from that code are a secondary consideration.
It would be better to say that to a GPL advocate, the code is the application. They are one in the same. The code therefore deserves to be protected under the GPL for the sake of the application.
Most applications are dynamic. They grow and change over time, and if they stop growing they usually fall out of use. A person choosing the GPL for their application wants this growth to happen openly and freely. The author of this article/rant talks about applications released under BSD 15 years ago that are still free even though they are useless. Perhaps if they'd been under the GPL, some of those applications might have stayed alive and kept growing.
What is needed is a clear separation between those companies that sling bits (ISPs) and those who provide content and advertising. Each ISP should be required to transfer data as fairly as possible with a minimum of interference and monitoring.
Most broadband providers have a monopoly or duopoly, and therefore need to be regulated strongly. Otherwise, customers who object to these invasions of privacy will have nowhere to turn.
HD feeds wouldn't work out so well, as most people don't have the bandwidth. Bittorrent wouldn't be good for NBC, because think about what happens when the masses start learning how to use bittorrent. You know someone will edit out the commercials and create new torrents that will quickly become more popular than NBC's torrents.
However, it's not hopeless. I think they should create a proprietary, cross-platform P2P based Olympics viewer. A user could simply rank the events they were most interested in, and allow the program to download videos in the background. The program could then play these events back with a small number of ads included. The video ads should be brief, and supplemented by text ads on the user interface or below the video.
Most people don't care much about seeing the events live, so long as they don't know the outcome. Most events won't be taking place at the time people want to watch anyway.
There would be no need to encrypt the video. Let the determined few create torrents of the files, but the general public will accept a couple of ads in exchange for the convenience of using NBC's application, if it is well designed. Many people will also prefer the legal option, so long as NBC doesn't get greedy. Don't show 20 minutes of ads per hour like television, instead show 30 seconds every ten minutes. Otherwise you can't compete with bittorrent.
The Olympics is a global event, so the cost of recording the events should be shared among the various countries. Each country can then simply place their own commentators and graphics over the video feed. This will keep costs in check. (If they're really clever, they'll share the cost of developing the viewer application too.)
With globally shared video costs and P2P replacing broadcast costs, the expenses could be brought down so that the reduced number of ads still make a nice profit for NBC.
Time Warner's highest level tier for their experiment with usage-based pricing is 40 GB/month. This is less than the capacity of a single Blu-Ray disc. Sony must be doing a happy dance.
In Google's defense, they may have legitimate use for these records. Viewing history is clearly important if they want to offer better viewing suggestions to YouTube users. I also wonder if they include this data in their formula for presenting personalised Google search results.
I'm sure it would be possible for them to get by with only a month or two of records, but consider why it is that Google is so successful as a search engine. They go out of their way to use every source of data they can to optimize their search results. They're not going to just toss out a valuable source of information like this if they can help it.
The part that might not be possible is finding a cellular service provider to go along with it. Most plans seem to start at $40/month and go up from there. Pay as you go plans typically charge 20 cents per minute, which is great if you make an hour of calls per month, but if you make 5 hours of calls one month you're looking at $60 for that month. And if you want any data or text messages it complicates it further.
The advantage of using wi-fi for internet based calls is that by only making a fraction of my calls using cellular, I should only have to pay a fraction of the cellular fees that I would otherwise pay. Unfortunately, the cell providers have priced their plans to make this difficult. You only get the cheaper per-minute and per-byte rates if you buy a high-priced plan and use it often.
Hear, hear! I'd like to see the breakdown of what it would actually cost to use this on a U.S. network and what kind of service to expect.
I did find this list of U.S. GSM providers by following a link from OpenMoko's Q&A page, but it's not all that helpful.
Since this device has wi-fi, what I'd really like to do is to use it as a Skype-like phone over wi-fi when I'm at home, and then have it switch to regular cellular when I'm on the road. Since I make most of my phone calls from home, I should theoretically save a lot of money this way. Realistically, I don't know if this is possible.
There are two reasons this is not reasonable. For one thing, the majority of Netflix users are not using some portion of the service. "I don't have a Blu Ray player, so charge me less. I don't use Profiles, so charge me less." That would quickly become a mess. Secondly, if they offered you the discount you wanted, suddenly 80% of Netflix users would be claiming to be Linux or Mac users.
Anyway, Watch It Now is more of an R&D feature than a selling point. The overall price for Netflix service has actually gone down since it came out.
How many times is a disc rented out over its lifetime? That $10 extra per disc gets divided up quite a bit.
Furthermore, why should BD users get singled out to be charged higher rates? Users who constantly send their discs back within a day don't get charged more. Users who use the profile feature (which we now know is expensive to maintain) don't get charged more. Users who use the "Watch It Now" online streaming don't get charged more, although that must be a hugely expensive experiment. So why single out BD users? If they really are a small percentage of the users, then the extra operating expense will also be relatively small.
I saw an optometrist recently who explained that because of the way pixels are displayed on a monitor, our eyes are constantly refocusing while we're at the computer. This can cause a lot of strain over time. There are glasses available which are designed specifically for computer use which can alleviate this problem.
There is a warning sign that your eyes are overstrained. If you move your eyes from focusing close to focusing on something far away, and you experience a delay in focusing, that means trouble. It's time to turn off the monitor and make an appointment with your eye doctor.
In what sense is this a robot? As I understand it, it moves back and forth along your gutter when you push the corresponding buttons. If this is a robot, then so is an RC car.
A gutter cleaning robot would emerge from a pod at the top of your roof, and walk around with spider-like legs. It would first map out your roof and send a 3D map to your computer. You would then indicate on the map where it should dump the gutter contents over the edge, and it would go to work. It might take all day to clear out the gutters, but you would be free to leave it alone (after you got bored of watching it.)
I don't get it either. It strikes me as security theater, or simply a means to harass people they don't like. What's clear is that the motivation is NOT for the common good, and so these searches are simply indefensible.
If there is evil, harmful data coming through our borders on laptops, then surely there is also evil data in people's minds. Should we find a way to search those too?
I hear that AI is only about 15 years away, so you could try just waiting until then. Unfortunately, that estimate hasn't changed for 30 years.
Given the slow progress in AI research, I think a radical approach is in order. I doubt that we'll see any breakthroughs from a small crew of programmers with quad cores and c++.
The human brain is a massively parallel, self-reconfiguring network of nodes. How far have we come in building any sort of scalable technology that can operate in such a manner? I know there are projects to try and reverse engineer brains from creatures of various intelligences. But even if we succeed in getting the basic blueprint of a simple brain, how would we go about building it? Because of this, I would lean towards funding projects that are developing new kinds of hardware inspired by the brain's design. Without currently unheard-of levels of parallelization, advanced AI may not even be physically possible.
If you haven't read "Surely You're Joking, Mr. Feynman!", I suggest you look it up. You'll enjoy his stories of how he taught himself to fix radios as a teen and made a business out of it.
(In fact, I have two copies because I received one as a gift. I'll send you a copy for no charge if you want it. Just, um, reply here I guess.)
Slashdot Mirror
Technically you're right. But I'm pretty sure that if some idiot chose "me@mydomain.com" as his personal email address, he's already used to getting mountains of spam.
There is no single Condorcet method. There are several Condorcet methods with various pros and cons, but I think Ranked Pairs manages to be the most fair. It's certainly much fairer than IRV.
Calculating the Ranked Pairs winner can be complicated, but voting is easy. Simply rank the candidates in order of preference, and skip those you don't know/care about. Anyone can follow that.
Approval voting has a nice simplicity to it, but it fails to distinguish between strong approval and weak approval, and I think it creates much more temptation for people to game the system by how they vote.
Whatever the system chosen, it would be a huge help in changing politics for the better. Unfortunately, it's not likely to happen in our lifetimes.
Ok, you caught me. But the first person to run with it ought to be able to pick up a few million in VC money.
It would seem that Ars Technica doesn't know what a Google "20 percent project" is.
This is the first in O'Reilly's new Buzzword series. Coming soon: "Blogospheres", "Synergy", and "Social Cloudcasting"
It would be better to say that to a GPL advocate, the code is the application. They are one in the same. The code therefore deserves to be protected under the GPL for the sake of the application.
Most applications are dynamic. They grow and change over time, and if they stop growing they usually fall out of use. A person choosing the GPL for their application wants this growth to happen openly and freely. The author of this article/rant talks about applications released under BSD 15 years ago that are still free even though they are useless. Perhaps if they'd been under the GPL, some of those applications might have stayed alive and kept growing.
Tag this story "dpwned".
What is needed is a clear separation between those companies that sling bits (ISPs) and those who provide content and advertising. Each ISP should be required to transfer data as fairly as possible with a minimum of interference and monitoring.
Most broadband providers have a monopoly or duopoly, and therefore need to be regulated strongly. Otherwise, customers who object to these invasions of privacy will have nowhere to turn.
HD feeds wouldn't work out so well, as most people don't have the bandwidth. Bittorrent wouldn't be good for NBC, because think about what happens when the masses start learning how to use bittorrent. You know someone will edit out the commercials and create new torrents that will quickly become more popular than NBC's torrents.
However, it's not hopeless. I think they should create a proprietary, cross-platform P2P based Olympics viewer. A user could simply rank the events they were most interested in, and allow the program to download videos in the background. The program could then play these events back with a small number of ads included. The video ads should be brief, and supplemented by text ads on the user interface or below the video.
Most people don't care much about seeing the events live, so long as they don't know the outcome. Most events won't be taking place at the time people want to watch anyway.
There would be no need to encrypt the video. Let the determined few create torrents of the files, but the general public will accept a couple of ads in exchange for the convenience of using NBC's application, if it is well designed. Many people will also prefer the legal option, so long as NBC doesn't get greedy. Don't show 20 minutes of ads per hour like television, instead show 30 seconds every ten minutes. Otherwise you can't compete with bittorrent.
The Olympics is a global event, so the cost of recording the events should be shared among the various countries. Each country can then simply place their own commentators and graphics over the video feed. This will keep costs in check. (If they're really clever, they'll share the cost of developing the viewer application too.)
With globally shared video costs and P2P replacing broadcast costs, the expenses could be brought down so that the reduced number of ads still make a nice profit for NBC.
It may never happen, but one can dream...
So you're saying that the yellowcake is a lie?
Time Warner's highest level tier for their experiment with usage-based pricing is 40 GB/month. This is less than the capacity of a single Blu-Ray disc. Sony must be doing a happy dance.
In Google's defense, they may have legitimate use for these records. Viewing history is clearly important if they want to offer better viewing suggestions to YouTube users. I also wonder if they include this data in their formula for presenting personalised Google search results.
I'm sure it would be possible for them to get by with only a month or two of records, but consider why it is that Google is so successful as a search engine. They go out of their way to use every source of data they can to optimize their search results. They're not going to just toss out a valuable source of information like this if they can help it.
The part that might not be possible is finding a cellular service provider to go along with it. Most plans seem to start at $40/month and go up from there. Pay as you go plans typically charge 20 cents per minute, which is great if you make an hour of calls per month, but if you make 5 hours of calls one month you're looking at $60 for that month. And if you want any data or text messages it complicates it further.
The advantage of using wi-fi for internet based calls is that by only making a fraction of my calls using cellular, I should only have to pay a fraction of the cellular fees that I would otherwise pay. Unfortunately, the cell providers have priced their plans to make this difficult. You only get the cheaper per-minute and per-byte rates if you buy a high-priced plan and use it often.
Hear, hear! I'd like to see the breakdown of what it would actually cost to use this on a U.S. network and what kind of service to expect.
I did find this list of U.S. GSM providers by following a link from OpenMoko's Q&A page, but it's not all that helpful.
Since this device has wi-fi, what I'd really like to do is to use it as a Skype-like phone over wi-fi when I'm at home, and then have it switch to regular cellular when I'm on the road. Since I make most of my phone calls from home, I should theoretically save a lot of money this way. Realistically, I don't know if this is possible.
There are two reasons this is not reasonable. For one thing, the majority of Netflix users are not using some portion of the service. "I don't have a Blu Ray player, so charge me less. I don't use Profiles, so charge me less." That would quickly become a mess. Secondly, if they offered you the discount you wanted, suddenly 80% of Netflix users would be claiming to be Linux or Mac users.
Anyway, Watch It Now is more of an R&D feature than a selling point. The overall price for Netflix service has actually gone down since it came out.
How many times is a disc rented out over its lifetime? That $10 extra per disc gets divided up quite a bit.
Furthermore, why should BD users get singled out to be charged higher rates? Users who constantly send their discs back within a day don't get charged more. Users who use the profile feature (which we now know is expensive to maintain) don't get charged more. Users who use the "Watch It Now" online streaming don't get charged more, although that must be a hugely expensive experiment. So why single out BD users? If they really are a small percentage of the users, then the extra operating expense will also be relatively small.
I bet this was fun to see. Did he quite literally fly backwards out of his office chair?
It's rare in the Army to be quite literally blown away and live to tell about it.
hmmmmm, you're right. this could get out of control.
hmmmmmm. why did you start this?
I saw an optometrist recently who explained that because of the way pixels are displayed on a monitor, our eyes are constantly refocusing while we're at the computer. This can cause a lot of strain over time. There are glasses available which are designed specifically for computer use which can alleviate this problem.
There is a warning sign that your eyes are overstrained. If you move your eyes from focusing close to focusing on something far away, and you experience a delay in focusing, that means trouble. It's time to turn off the monitor and make an appointment with your eye doctor.
In what sense is this a robot? As I understand it, it moves back and forth along your gutter when you push the corresponding buttons. If this is a robot, then so is an RC car.
A gutter cleaning robot would emerge from a pod at the top of your roof, and walk around with spider-like legs. It would first map out your roof and send a 3D map to your computer. You would then indicate on the map where it should dump the gutter contents over the edge, and it would go to work. It might take all day to clear out the gutters, but you would be free to leave it alone (after you got bored of watching it.)
Actually, I believe the date, because of this quote: "our approach with Windows 7 is to build off the same core architecture as Windows Vista."
Translation: Windows 7 will be Vista SP2 with a few shiny bits attached. It will be on time, and pointless.
I don't get it either. It strikes me as security theater, or simply a means to harass people they don't like. What's clear is that the motivation is NOT for the common good, and so these searches are simply indefensible.
If there is evil, harmful data coming through our borders on laptops, then surely there is also evil data in people's minds. Should we find a way to search those too?
This is the problem with the open source world. Try spelling like that at Microsoft, and someone is likely to throw a chare.
I hear that AI is only about 15 years away, so you could try just waiting until then. Unfortunately, that estimate hasn't changed for 30 years.
Given the slow progress in AI research, I think a radical approach is in order. I doubt that we'll see any breakthroughs from a small crew of programmers with quad cores and c++.
The human brain is a massively parallel, self-reconfiguring network of nodes. How far have we come in building any sort of scalable technology that can operate in such a manner? I know there are projects to try and reverse engineer brains from creatures of various intelligences. But even if we succeed in getting the basic blueprint of a simple brain, how would we go about building it? Because of this, I would lean towards funding projects that are developing new kinds of hardware inspired by the brain's design. Without currently unheard-of levels of parallelization, advanced AI may not even be physically possible.
If you haven't read "Surely You're Joking, Mr. Feynman!", I suggest you look it up. You'll enjoy his stories of how he taught himself to fix radios as a teen and made a business out of it.
(In fact, I have two copies because I received one as a gift. I'll send you a copy for no charge if you want it. Just, um, reply here I guess.)