Very true, which is why most companies that sell GPL software either bundle it with hardware and sell it as an appliance (i.e. many NAS manufacturers, Barracuda Networks, Watchguard, Linksys, TiVo, etc.), or they sell it with support (i.e. Red Hat, Novell, etc.). The sad thing about the ones that sell appliances is that many of them do not make it very apparent that they are running open source, and just include the required GPL notice and method of obtaining source on a slip of paper in the box. Which gets back to my earlier statement that larger GPL adoption will lead to less direct software sales and will lead to a larger service industry for software.
I never stated it was stealing nor that there was a problem, hence the "This is fine if all you want to do is give away your code." Hell, you even included that part when you quoted me. The post I replied to was stating that programmers should release their work as LGPL or BSD to broaden adoption. I was stating that there are very valid reasons not to. By pushing BSD-style licensing, there is little incentive to keep the code open and derivative works can be sold closed-source. By pushing GPL-style licensing, closed-source companies are forced to either make separate licensing agreements with the creator, or they need to follow the GPL and distribute their modified source.
My personal opinion on what open source license should everyone use? Whatever they want. It is their code, and no-one should force them to release it under a license they don't agree with. Now can both sides stop the bickering and bitching about which license is best? They both serve a purpose, and the in-fighting harms the image of open source in general.
If you have to distribute the source-code of your derivative work under the GPL, I fail not see how they are "proprietary" any longer. You are correct in that you can charge for your derivative works. However, you cannot sell them under a different license.
How many cubic handbreadths of dark ale equals one arse-load again? I have 68 furlongs to drive home and I need to know how much Guinness to bring. I know I can drink about a third of an arse-load in 26 furlongs, but the stupid store here sells by the cubic handbreadth and my phone does not have the proper conversion tools. Damn technology does everything in the world as long as it isn't useful.
It all depends on where you want to steer open-source adoption to.
With BSD-style, you are mostly steering open-source adoption to software houses, who will use the code to save money on initial development and sell proprietary versions of the code (i.e. the Microsoft TCP/IP stack example that you mentioned). You are basically giving your code away. A major software house can take this code, modify it, resell it and not have to give back the modifications. This is fine if all you want to do is give away your code.
GPL is more of a "I'll scratch your back, but only if you scratch mine". In order for a company to take GPL code, modify it, and resell it, they need to give their modifications back. With the GPL, less money will be made in selling the software itself, and more will need to be made in other areas, like support. This means that there will be less big software companies like Microsoft, Symantec, and EMC. And there will be more consulting companies with programmers on staff customizing and deploying software for the clients.
I believe this depends on the state as well. In my state it is basically "hire at will". An employee can be basically not hired or fired as long as there is no obvious illegal discrimination. However, most companies usually wait until there is a sizable amount of documented disciplinary actions before firing. That way the ex-employee will have a much more difficult time of getting unemployment benefits.
I've heard that some states are very different and that there are more employee-favored regulations, such as mandatory 2-week notices before firing an employee, or employee must receive their final pay check within 24 hours of termination.
At the moment, the King's edict is that the imperial fuckload is larger. However, a metric shit-tonne remains larger still. You could pour 4,516 gill of treacle into a barrel and it would only weigh about half a shit-tonne.
I'd be willing to bet that the main reason that torrents are coming from cable tv rips is not because of the security. Usually cable gets the content before the streaming networks. It is also often of higher quality than streaming video. However, all of these things being equal (release time, quality), cable tv rips will probably still prevail due to the security.
tl;dr version: The security is mostly unnecessary as everyone just rips the videos from cable since it is out sooner and is better quality. Putting so many resources into securing the web streams is like locking the barn door after the horse escapes.
The injunction is terrible in that it was not lifted during appeals. That paints a scary picture for manufacturers as they cannot even keep their business going during appeals. If it weren't for their lucrative storage products, Buffalo Tech likely would have had to close the US office.
The funny thing is that the CSIRO patent in question specifically speaks of using frequencies above 10 GHz. Since 802.11G uses 2.4 Ghz and A uses 5 GHz, how is this covered by the patent?
My question was rhetorical in that I was stating that the manufacturers pretty much do the same thing. In fact, I don't think Buffalo even owns a chip fab to make the chips that Broadcom, etc. design.
Maybe they shouldn't be able to change the frequencies mentioned in the patent. Maybe they shouldn't be able to patent obvious things like multiplexing over different frequencies. Maybe they shouldn't bring the lawsuit up in a court district that is a known haven to patent trolls. Maybe they should go after the WIFI Alliance, who knowingly introduced the standard including the patented tech, but did not provision for licensing fees. On that subject, who should be responsible for those fees? The WIFI Alliance? The chip designers (Broadcom, RALink, etc.)? The manufacturers (Netgear, Buffalo, etc.)? The companies and individuals using the equipment?
A standard should not include patented technology. Or, the patent holder should be contractually obliged to not sue over the patents when they are used within the standard before the standard can be introduced.
Another problem is who should be responsible for paying the license? I'm pretty sure Buffalo(among others) does not currently design its own chips. They are designed by companies like Broadcom and Atheros. Since Broadcom and Atheros are designing chips that use the patents, shouldn't they bear the cost of licensing? The parts of the products that are designed by Buffalo shouldn't really infringe on the patents directly, as they just use the chips that themselves use the patented technology. If I buy a computer and a wireless adapter, install linux and configure it as an access point or router, and then sell the wireless access point to another person, should I have to pay licenses to wifi patent owners?
A scammer is a scammer, whether they are running back-alley games of 3 card Monte or are distributing applications loaded with spyware or other malware. Just because it is ignorance causing people to fall for these scams does not mean the scammers should get away with it.
One might as well say that they shouldn't go after people that rob houses because the houses should have had better locks.
Seriously though, even if Windows 7 was the most secure OS out there, undereducated users are the problem. Not enough people are afraid to install any old thing they find on the internet. Even without standard exploits, trojans work because the user chooses to install them. There is no way to stop that is with user education or by preventing users from installing anything other than vetted software. Most users are too lazy for the former and Microsoft cannot do the latter without risking yet another antitrust lawsuit. A linux distro with a good package manager is a nice step because most software a user would need can be easily found and installed from a trusted source. However, a trojan can still be packaged into FREESMILIES.deb and can be installed with a double-click in nearly any distro with apt, sometimes easier than a.exe in windows. The only way to stop that in linux is to force packages to be only installed from the trusted repositories, or make it so difficult to install untrusted packages that the average joe would not bother.
Thanks for the info. I was basing my info for IE8 off of something I read back when it was announced that they were changing the architecture, and I distinctly remember that they said threads and not processes. Separate processes makes more sense, so I always thought it was pretty strange, but never cared enough about IE to check.
The original intent of Starter Edition was to be an ultra-cheap version to combat piracy in "undeveloped" nations. Now, it is being pushed worldwide to be an ultra-cheap version for netbook computers so that manufacturers are less inclined to use linux. They may say the limits are because the hardware is weaker, but in that case they should have released a low-footprint version instead of crippling the user. Granted, a low-footprint version would not include unnecessary services and would be crippled in a sense, but the user should be able to decide how many simultaneous apps they want to run. There is some stripped down version of XP that came out a while back that was called something like "Windows Essentials for Legacy Computers". I'm pretty sure that is more in line with a low-footprint version than the "Starter Edition".
A better car analogy would be a cheap van with 10 seats but it will not allow more than 3 passengers. If you pay to upgrade(just removes the limiter[3 app limit], adds a DVD player[Media Center], and some decorations[Aero glass]), you can have as many passengers as you can fit.
A virtual machine should only count as one "application", and any processes running on the guest should not be visible to the host OS. Hell, VMWare server runs as a background service, and the console is a browser plug-in. So you can still use the browser for whatever, and have a window open for your VM console, and all of that is only 1 "application".
Then again, why do that? Why not just install linux and run WINE. Having a useless host OS is just a bunch of unnecessary overhead.
You've probably never worked on somebody else's netbook before. Many people just try to use them like they would a normal computer(i.e. install everything they see and try to run 50 apps at once). If you think tons of toolbars look bad on a regular screen, imagine that on a 7 inch netbook screen.
Other users have posted this, and it gives some good information as to what is counted and what is not. Something that runs in the tray ("notification area") does not seem to count, but may when the window is launched. An example from the article is an antivirus app that runs in the tray. While in the tray, it works fine, but if you launch the main window for the application that window counts toward the limit. This means you might be able to hide applications in the tray using TrayIt or something, but only 3 can be unhidden at a time.
Nvidia and other chip manufacturers still do this and it is all perfectly legal. Normally the clocked down and reduced-core chips come from the defect bin. A good example is the 3-core processor from AMD. It is the exact same as the 4 core model, but one of the cores is deactivated. Usually this is done because one of the cores is bad, but sometimes if there is a higher demand on the cheaper product, they may limit perfectly fine products to increase supply. Another example, and possibly the one you are referring to is the Nvidia GeForce 6200 and 6600. Earlier models of the 6200 were 6600 chips with a different firmware, normally the ones from the defect bin. This is part of the reason 6200 cards suck terribly. Then again, you could "softmod" some of these 6200 up to a 6600 and get a free upgrade. You may also be referring to the differences between the GeForce and Quadro lines. Many of the Quadro chips are identical to Geforce chips but with a different firmware. Correct me if I am wrong, but it is my understanding that the Quadro line is not actually faster, but performs rendering differently. In 3d design, rendering accuracy is much more important than speed, so Quadro is optimized for accuracy and they charge a hefty price for that optimization.
There is an article over on ars technica that goes into detail about the problems you are describing. It actually compares OSX and Windows, though. Starter Edition is not basing its "application" count on number of windows open, otherwise 3 word docs would be the limit. It is not basing it on processes, otherwise the damn thing wouldn't boot or would need a "base system" whitelist. It seems to be basing it on unique desktop applications, the same as the Applications tab in Task Manager. It does appear to have a whitelist for Windows utilities as Explorer and cmd windows do not count toward your limit. Also, applications with multiple processes work fine, like Chrome and its process-per-tab architecture.
Not a single one of those is considered an "application", those are services. Applications are pretty much defined as what shows up under "Applications" in task manager minus most built in utilities (i.e. explorer, control panel, installers, etc.). Since VMWare server runs as a background service, there should be no problems. In fact, the console for your VM runs in a browser window, so your web browsing plus VMs only count as 1 application. Then again, if you are going to go through that much work to circumvent this limit, just install Ubuntu netbook remix or another linux environment. In a netbook you probably don't need any windows apps that don't have decent linux replacements or versions.
Well, you completely missed the joke there, but I'll try to make helpful response. Chrome uses one process per tab. IE8 uses a single process and separates tabs into threads of the process. It seems that Starter Edition is basing this on the number of "applications" based on what shows up under "Applications" tab in Task Manager. So Chrome and IE should each only be counted as a single application toward the limit.
Someone else mentioned using tray apps that hide applications in the system tray instead of in the task bar to circumvent the limitation. I suspect that might work if it also removes it from the "Applications" tab of Task Manager.
Slashdot Mirror
Very true, which is why most companies that sell GPL software either bundle it with hardware and sell it as an appliance (i.e. many NAS manufacturers, Barracuda Networks, Watchguard, Linksys, TiVo, etc.), or they sell it with support (i.e. Red Hat, Novell, etc.). The sad thing about the ones that sell appliances is that many of them do not make it very apparent that they are running open source, and just include the required GPL notice and method of obtaining source on a slip of paper in the box. Which gets back to my earlier statement that larger GPL adoption will lead to less direct software sales and will lead to a larger service industry for software.
I never stated it was stealing nor that there was a problem, hence the "This is fine if all you want to do is give away your code." Hell, you even included that part when you quoted me. The post I replied to was stating that programmers should release their work as LGPL or BSD to broaden adoption. I was stating that there are very valid reasons not to. By pushing BSD-style licensing, there is little incentive to keep the code open and derivative works can be sold closed-source. By pushing GPL-style licensing, closed-source companies are forced to either make separate licensing agreements with the creator, or they need to follow the GPL and distribute their modified source.
My personal opinion on what open source license should everyone use? Whatever they want. It is their code, and no-one should force them to release it under a license they don't agree with. Now can both sides stop the bickering and bitching about which license is best? They both serve a purpose, and the in-fighting harms the image of open source in general.
If you have to distribute the source-code of your derivative work under the GPL, I fail not see how they are "proprietary" any longer. You are correct in that you can charge for your derivative works. However, you cannot sell them under a different license.
How many cubic handbreadths of dark ale equals one arse-load again? I have 68 furlongs to drive home and I need to know how much Guinness to bring. I know I can drink about a third of an arse-load in 26 furlongs, but the stupid store here sells by the cubic handbreadth and my phone does not have the proper conversion tools. Damn technology does everything in the world as long as it isn't useful.
It all depends on where you want to steer open-source adoption to.
With BSD-style, you are mostly steering open-source adoption to software houses, who will use the code to save money on initial development and sell proprietary versions of the code (i.e. the Microsoft TCP/IP stack example that you mentioned). You are basically giving your code away. A major software house can take this code, modify it, resell it and not have to give back the modifications. This is fine if all you want to do is give away your code.
GPL is more of a "I'll scratch your back, but only if you scratch mine". In order for a company to take GPL code, modify it, and resell it, they need to give their modifications back. With the GPL, less money will be made in selling the software itself, and more will need to be made in other areas, like support. This means that there will be less big software companies like Microsoft, Symantec, and EMC. And there will be more consulting companies with programmers on staff customizing and deploying software for the clients.
I believe this depends on the state as well. In my state it is basically "hire at will". An employee can be basically not hired or fired as long as there is no obvious illegal discrimination. However, most companies usually wait until there is a sizable amount of documented disciplinary actions before firing. That way the ex-employee will have a much more difficult time of getting unemployment benefits.
I've heard that some states are very different and that there are more employee-favored regulations, such as mandatory 2-week notices before firing an employee, or employee must receive their final pay check within 24 hours of termination.
At the moment, the King's edict is that the imperial fuckload is larger. However, a metric shit-tonne remains larger still. You could pour 4,516 gill of treacle into a barrel and it would only weigh about half a shit-tonne.
I'd be willing to bet that the main reason that torrents are coming from cable tv rips is not because of the security. Usually cable gets the content before the streaming networks. It is also often of higher quality than streaming video. However, all of these things being equal (release time, quality), cable tv rips will probably still prevail due to the security.
tl;dr version: The security is mostly unnecessary as everyone just rips the videos from cable since it is out sooner and is better quality. Putting so many resources into securing the web streams is like locking the barn door after the horse escapes.
The injunction is terrible in that it was not lifted during appeals. That paints a scary picture for manufacturers as they cannot even keep their business going during appeals. If it weren't for their lucrative storage products, Buffalo Tech likely would have had to close the US office.
The funny thing is that the CSIRO patent in question specifically speaks of using frequencies above 10 GHz. Since 802.11G uses 2.4 Ghz and A uses 5 GHz, how is this covered by the patent?
My question was rhetorical in that I was stating that the manufacturers pretty much do the same thing. In fact, I don't think Buffalo even owns a chip fab to make the chips that Broadcom, etc. design.
I guess I should stop giving patent holders amnesia after "stealing" their designs, then.
Maybe they shouldn't be able to change the frequencies mentioned in the patent. Maybe they shouldn't be able to patent obvious things like multiplexing over different frequencies. Maybe they shouldn't bring the lawsuit up in a court district that is a known haven to patent trolls. Maybe they should go after the WIFI Alliance, who knowingly introduced the standard including the patented tech, but did not provision for licensing fees. On that subject, who should be responsible for those fees? The WIFI Alliance? The chip designers (Broadcom, RALink, etc.)? The manufacturers (Netgear, Buffalo, etc.)? The companies and individuals using the equipment?
A standard should not include patented technology. Or, the patent holder should be contractually obliged to not sue over the patents when they are used within the standard before the standard can be introduced.
Another problem is who should be responsible for paying the license? I'm pretty sure Buffalo(among others) does not currently design its own chips. They are designed by companies like Broadcom and Atheros. Since Broadcom and Atheros are designing chips that use the patents, shouldn't they bear the cost of licensing? The parts of the products that are designed by Buffalo shouldn't really infringe on the patents directly, as they just use the chips that themselves use the patented technology. If I buy a computer and a wireless adapter, install linux and configure it as an access point or router, and then sell the wireless access point to another person, should I have to pay licenses to wifi patent owners?
A scammer is a scammer, whether they are running back-alley games of 3 card Monte or are distributing applications loaded with spyware or other malware. Just because it is ignorance causing people to fall for these scams does not mean the scammers should get away with it.
One might as well say that they shouldn't go after people that rob houses because the houses should have had better locks.
There is no way to stop that is with user education or by preventing users from installing anything other than vetted software.
Massive brain fart while typing that sentence. Should read "The only way to stop that is with..."
Time for more coffee.
He did say reformat AFTER upgrading to Windows 7.
Seriously though, even if Windows 7 was the most secure OS out there, undereducated users are the problem. Not enough people are afraid to install any old thing they find on the internet. Even without standard exploits, trojans work because the user chooses to install them. There is no way to stop that is with user education or by preventing users from installing anything other than vetted software. Most users are too lazy for the former and Microsoft cannot do the latter without risking yet another antitrust lawsuit. A linux distro with a good package manager is a nice step because most software a user would need can be easily found and installed from a trusted source. However, a trojan can still be packaged into FREESMILIES.deb and can be installed with a double-click in nearly any distro with apt, sometimes easier than a .exe in windows. The only way to stop that in linux is to force packages to be only installed from the trusted repositories, or make it so difficult to install untrusted packages that the average joe would not bother.
Thanks for the info. I was basing my info for IE8 off of something I read back when it was announced that they were changing the architecture, and I distinctly remember that they said threads and not processes. Separate processes makes more sense, so I always thought it was pretty strange, but never cared enough about IE to check.
The original intent of Starter Edition was to be an ultra-cheap version to combat piracy in "undeveloped" nations. Now, it is being pushed worldwide to be an ultra-cheap version for netbook computers so that manufacturers are less inclined to use linux. They may say the limits are because the hardware is weaker, but in that case they should have released a low-footprint version instead of crippling the user. Granted, a low-footprint version would not include unnecessary services and would be crippled in a sense, but the user should be able to decide how many simultaneous apps they want to run. There is some stripped down version of XP that came out a while back that was called something like "Windows Essentials for Legacy Computers". I'm pretty sure that is more in line with a low-footprint version than the "Starter Edition".
A better car analogy would be a cheap van with 10 seats but it will not allow more than 3 passengers. If you pay to upgrade(just removes the limiter[3 app limit], adds a DVD player[Media Center], and some decorations[Aero glass]), you can have as many passengers as you can fit.
Then again, if you get infected, and the malware spawns 3 applications, you wont be able to run whatever application you are trying to use to fix it.
A virtual machine should only count as one "application", and any processes running on the guest should not be visible to the host OS. Hell, VMWare server runs as a background service, and the console is a browser plug-in. So you can still use the browser for whatever, and have a window open for your VM console, and all of that is only 1 "application".
Then again, why do that? Why not just install linux and run WINE. Having a useless host OS is just a bunch of unnecessary overhead.
You've probably never worked on somebody else's netbook before. Many people just try to use them like they would a normal computer(i.e. install everything they see and try to run 50 apps at once). If you think tons of toolbars look bad on a regular screen, imagine that on a 7 inch netbook screen.
http://blogs.zdnet.com/Bott/?p=844
Other users have posted this, and it gives some good information as to what is counted and what is not. Something that runs in the tray ("notification area") does not seem to count, but may when the window is launched. An example from the article is an antivirus app that runs in the tray. While in the tray, it works fine, but if you launch the main window for the application that window counts toward the limit. This means you might be able to hide applications in the tray using TrayIt or something, but only 3 can be unhidden at a time.
Nvidia and other chip manufacturers still do this and it is all perfectly legal. Normally the clocked down and reduced-core chips come from the defect bin. A good example is the 3-core processor from AMD. It is the exact same as the 4 core model, but one of the cores is deactivated. Usually this is done because one of the cores is bad, but sometimes if there is a higher demand on the cheaper product, they may limit perfectly fine products to increase supply. Another example, and possibly the one you are referring to is the Nvidia GeForce 6200 and 6600. Earlier models of the 6200 were 6600 chips with a different firmware, normally the ones from the defect bin. This is part of the reason 6200 cards suck terribly. Then again, you could "softmod" some of these 6200 up to a 6600 and get a free upgrade. You may also be referring to the differences between the GeForce and Quadro lines. Many of the Quadro chips are identical to Geforce chips but with a different firmware. Correct me if I am wrong, but it is my understanding that the Quadro line is not actually faster, but performs rendering differently. In 3d design, rendering accuracy is much more important than speed, so Quadro is optimized for accuracy and they charge a hefty price for that optimization.
There is an article over on ars technica that goes into detail about the problems you are describing. It actually compares OSX and Windows, though. Starter Edition is not basing its "application" count on number of windows open, otherwise 3 word docs would be the limit. It is not basing it on processes, otherwise the damn thing wouldn't boot or would need a "base system" whitelist. It seems to be basing it on unique desktop applications, the same as the Applications tab in Task Manager. It does appear to have a whitelist for Windows utilities as Explorer and cmd windows do not count toward your limit. Also, applications with multiple processes work fine, like Chrome and its process-per-tab architecture.
Not a single one of those is considered an "application", those are services. Applications are pretty much defined as what shows up under "Applications" in task manager minus most built in utilities (i.e. explorer, control panel, installers, etc.). Since VMWare server runs as a background service, there should be no problems. In fact, the console for your VM runs in a browser window, so your web browsing plus VMs only count as 1 application. Then again, if you are going to go through that much work to circumvent this limit, just install Ubuntu netbook remix or another linux environment. In a netbook you probably don't need any windows apps that don't have decent linux replacements or versions.
Well, you completely missed the joke there, but I'll try to make helpful response. Chrome uses one process per tab. IE8 uses a single process and separates tabs into threads of the process. It seems that Starter Edition is basing this on the number of "applications" based on what shows up under "Applications" tab in Task Manager. So Chrome and IE should each only be counted as a single application toward the limit.
Someone else mentioned using tray apps that hide applications in the system tray instead of in the task bar to circumvent the limitation. I suspect that might work if it also removes it from the "Applications" tab of Task Manager.