The FSF is claiming to be the good guys here, so why should they do something so underhanded as to claim a new license alters previous agreements retro-actively?
The FSF is doing no such thing.
GPLv3 is simply new terms on distributing FSF software in the future. There may be no way for Microsoft to honor the vouchers they've distributed, avoid the new GPLv3 terms, *and* not infringe copyright - but that doesn't mean the FSF is altering any agreements retroactively, they're simply changing the terms of the license that they are offering their future work under. Microsoft shouldn't have assumed that the FSF would continue to offer their new software under GPLv2.
Well it means that a full crack of BD+ will require crackers to implement a virtual machine which acts in exactly the same way as the hardware VM would act.
That sounds like a fun project. It'd be a good bit of work, but if it wasn't then it wouldn't be anywhere near as fun.
In this case, you have to come up with something which can determine the full dynamic runtime execution path of a static binary - a currently unsolved problem in Computer Science, despite numerous attempts to do such a thing by some of the world's brightest minds.
I think this ends up being the halting problem. It's impossible in the general case, and really easy in a number of specific cases. In any case, solving it that way is probably significantly harder than just writing an emulator and letting it do it's thing.
We still can't mathematically prove that ciphers are unbreakable, but that doesn't mean that a modern cipher like AES is going to be broken. We absolutely do know enough about the math to say that it can't be broken simply by brute force.
I agree that worldwide standardization on a sub-par distro (Say SLED) wouldn't be great, but I think that brand dilution is a bigger problem for "Linux on the desktop" than distro consolidation is at the moment.
Having some variety is good, but the "Linux Market" has more than enough variety. If some consolidation occurred, that would probably be a good thing. In the Desktop / Server / Workstation space, we have Red Hat, Ubuntu, Debian, Gentoo and a lot of distros that are blatantly redundant to one of those. The right to fork exists, so there's a huge advantage to delaying such a fork for as long as possible by contributing to a major distro.
You don't seem to have understood the suggestion properly.
Consider a Firefox bugfix - every Ubuntu machine that is being used as a desktop will pop up an "update available" notice and allow the user to download the update. If the Ubuntu people count the number of downloads of *that update*, they'll get a reasonably accurate count of internet-connected Ubuntu desktops.
EVDO gets something in the range of 2Mbps--noticeably faster than EDGE--and it seems slow to me.
Unless you're downloading very large files, that's because of the latency rather than the bandwidth. Latency is horrible on *all* current cellular data standards, but it's not that much more horrible on the "2.5G" networks compared to the "3G" networks.
For those of you who really think that the difference between 2.5G and 3G is a deal breaker, I'd like to point something out:
Unless you are going to hook the phone up to your computer and use it as a modem, the difference doesn't matter. You've got a 2" screen with a relatively low resolution - even crappy video streaming will run over 2.5G (poorly, but who watches video on their phone anyway?).
If you could do VOIP or something over 3G that might make it matter, but the latency for cellular internet access is so horrible that it's not worth it. On my 3G Sprint (PowerVision) phone, I've never seen my ping get lower than 500ms - and I've even written midlets to test it.
Maybe 4G will matter, but the difference between 2.5G and 3G is *nothing* relevant for any phone usage pattern I can come up with. Even MP3 downloads are fast enough on 2.5G.
There is a censoring system in place, and someone decides to use it for their own agenda.
The fact that the people in Sweden let it get that far makes me sad. It's blatantly obvious that once you give the authorities censorship power, they'll use it against groups they don't like. The question of "what are they censoring" shouldn't even enter into it - giving those tools to the state is a horrible mistake.
Using a phone to try to solve that sort of problem is correct only when the people you talk to on the phone have the authority to fix your problem. The minute you realize that they don't have sufficient authority, you need to switch to postal mail with signature confirmation.
A good old "Dear , you are billing me for a service I never actually requested. Here are the details:... Cease and desist immediately." on paper that they had to sign for will almost always produce the intended results. If it doesn't, at least you've got documentation to show a laywer.
Excuse me? I need to tailor my hardware purchases around my operating system? Are you nuts?
How is that nuts? The hardware is pretty useless if it doesn't run the software you want to use.
I expect Linux to run on the same machine where Windows XP will run; after all, in theory there shouldn't be a reason why this shouldn't occur without a problem.
No. In theory, as in practice, hardware only works when the OS supports it.
Windows doesn't support every piece of hardware that works with Linux. Why would you expect Linux to support every last piece of hardware that Windows does?
In that case, the hardware you bought to run Windows may not run any other OS. Sucks that you didn't plan ahead, but there's no way it's the fault of some other OS that it doesn't run on a machine designed specifically to run Windows and nothing else.
Perhaps, in the future, you'll consider software flexibility as one of your purchasing criteria.
You're right that C# and Java require super long columns to be even vaguely legible.
Personally, I'd say that's a flaw in the language. Java style interface naming, where the stock mutable string type is called "StringBuffer" (12 characters) is absurd, as is the idea that absolutely everything deserves a level of indentation (you can have "class { methood { lock { try {" before you even get to code). I've worked in a lot of programming languages, and none of them are perfect, but Java-like languages waste horizontal screen space like it was worthless.
Parent is correct. The only time a real user should see the OS is when it breaks (which it should not do). A fully wrapped OS (one you cant even see) is best for most users.
What you're saying is "some random idea that no-one's ever tried, much less successfully marketed is the one true answer". That's obviously utter hogwash. All of the desktop operating systems that have ever seen any adoption whatsoever have a certain amount of underlying structure that a power user or developer can get at relatively easily.
From this perspective, any modern OS (Windows Vista, Mac OS X, Ubuntu 7.04) are all fine. They hide enough that the average user can do what they want without worrying about the OS, and they reveal enough that the programmers can write software.
Adobe is an important player in the Linux on the Desktop story, but I don't think they're as significant as you make them out to be. They're software is way to expensive and blatantly unnecessary for most users, and both Sun and Microsoft have declared open season on their only really general-user product (flash).
I've been hearing this since I started using Linux back when I fell in love with Slackware 0.92. That version number may not be right, but it was around 1993. Linux is niche and it will always be niche until it has better gaming support and normal everyday office support.
You asked the person you replied to to back up his claims - please back up the claims I quoted here. Actually, don't bother - it's impossible, because the only thing that can possibly back up "Linux is nich and will always be niche..." is a gut feeling.
I'm IT for insurance. There is absolutely NOTHING available other than Windows based or mainframe based that I can run reliably.
This could be actually true, or it could be functionally true as the result of very conservative IT policies. Many industries have conservative IT policies, and for those industries change will come only very, very slowly.
But... here's the thing: Some specific systems in a conservative company's IT department have very little to do with overall market trends. Your use of Windows to track insurance will have precisely the same impact on Linux adoption that the mainframes had on Windows adoption in the early 90's - not much.
What's really important though is the fact that the computer industry isn't a single market. It's a whole bunch of smaller markets that all influence each-other to various degrees. And Linux is a solid, mainstream player everywhere from mainframes to cellphones. Feel free to bet that that won't translate into desktop gains, but that'd be a bet you'd lose.
Until you can take a distribution disk, pop it on a random machine with decent hardware, and have everything up and running without requiring any type of user action 'under the hood', Linux will remain firmly esconced in the realm of server rooms, geek basements, and nerd bedrooms; not in your average household.
Your expectation that an OS will work on random hardware makes no sense. You wouldn't expect Windows to run on a PPC Mac - why would you expect Ubuntu to give you 3D acceleration out of the box with an ATI Radeon X1300 (that has "Windows XP or higher" as a system requirement)?
When you buy a computer, you buy it for a specific purpose. You select hardware that can run the software you intend to use. The fact that some distribution of GNU/Linux doesn't provide perfect hardware support for the machine you bought to run Microsoft Windows shouldn't surprise you at all - in fact, every time you *succeed* at repurposing a machine in a manner that you didn't consider at the time of initial purchase you should be pleasantly surprised, since - in the great scheme of computer compatibility - that almost never happens.
Non-enthusiasts doing OS installs is a non-starter anyway. If your time is worth anything, you buy machines pre-installed - in which case any unsupported hardware means you return the computer as "broken". Dell sells Ubuntu desktops, as well as SLED and RHEL workstations.
Linux still has an enourmous amount of ground to cover before it comes close to being a serious rival to Windows in the consumer desktop market.
Again, the "Consumer Desktop Market" has NOTHING to do with people self-installing operating systems, much less self-installing an operating systems on systems that weren't specifically built to run that OS.
I would love a CPU that can power down to single-digits-of-watts in a low-power state.
You mean like a VIA C7?
Sure, we won't have the performance of today's 130 watt monsters in an 8 watt chip for a while, but you can sure get the performance of a slow Pentium IV in only a couple of watts.
Mips and Mflops are not useful benchmarks at all on modern processors. That measurement would be completely unaffected by any of the design decisions that effect actual code - stuff like cache and IO bandwidth just wouldn't matter at all. To use a bad car analogy, that's like asking what the sum of the theoretical maximum energy output per cycle of the cylinders - sure, it's a performance related metric, but not a useful one.
adding headers never improves speed. You're smoking crack.
Routing through a protocal-specific smart cache sometimes does though.
If an ISP doesn't want you to get to that one site, it is beyond trivial for them to block your "easy" attempts to get it.
Sure, the service needs to be offered at a finite number of IP addresses and the ISP can pretend to be a user to find out what they are, but even that method can be countered by silly techniques like user-partioning (different users see a different subset of servers based on a hash of their IP - so the ISP will only see 1/n of all of the servers). In the end, the best case for the ISP is that it turns into an arms race that eats endless man hours.
If you can find it easily, then they can block it easily.
That's false. Unless the ISP wants to move to destination address whitelisting (i.e. you can only access approved servers), there's no way to reliably prevent people from using encrypted VPN tunnels for their P2P traffic. If set up properly, such tunnels can actually *improve* P2P download speeds in some cases.
Intel CPU's now cost *less* than equivalent performing AMD CPU's
The best price/performance at the moment is the AMD 3600. From there, the price performance ratios get progressively worse as performance increases. Other processors that aren't blatantly obsolete include: The AMD 5000, Intel 4300, AMD 5600, and a bunch of Intel processors costing over $200.
If you don't mind, could you tell me if it would make any difference, overall, if the national debt were wiped out in America?
Yea, the national debit is about $5 trillion, so it would make a notifiable difference. The real question is this: What sort of difference would it make?
And if they can't figure out a.doc file, they probably won't be able to figure out opendocument any better, because it is just as silly to believe that opendocument will be any more common 1000 years from now than microsoft word documents.
The question isn't if the format will still be in use, the question is this: How hard will it be to write a converter to whatever the standard is in the future? With ODF, you need to uncompress a zip archive and parse some reasonably simple XML. With OOXML, you need to deal with all sorts of crap - and even if you have the standard, you won't necessarily be able to implement parts of it.
Slashdot Mirror
The FSF is doing no such thing.
GPLv3 is simply new terms on distributing FSF software in the future. There may be no way for Microsoft to honor the vouchers they've distributed, avoid the new GPLv3 terms, *and* not infringe copyright - but that doesn't mean the FSF is altering any agreements retroactively, they're simply changing the terms of the license that they are offering their future work under. Microsoft shouldn't have assumed that the FSF would continue to offer their new software under GPLv2.
That sounds like a fun project. It'd be a good bit of work, but if it wasn't then it wouldn't be anywhere near as fun.
I think this ends up being the halting problem. It's impossible in the general case, and really easy in a number of specific cases. In any case, solving it that way is probably significantly harder than just writing an emulator and letting it do it's thing.
Luckily, that stuff is available for free on "the internets". 720p in H.264 fits fine on 1 DVD.
We still can't mathematically prove that ciphers are unbreakable, but that doesn't mean that a modern cipher like AES is going to be broken. We absolutely do know enough about the math to say that it can't be broken simply by brute force.
I agree that worldwide standardization on a sub-par distro (Say SLED) wouldn't be great, but I think that brand dilution is a bigger problem for "Linux on the desktop" than distro consolidation is at the moment.
Having some variety is good, but the "Linux Market" has more than enough variety. If some consolidation occurred, that would probably be a good thing. In the Desktop / Server / Workstation space, we have Red Hat, Ubuntu, Debian, Gentoo and a lot of distros that are blatantly redundant to one of those. The right to fork exists, so there's a huge advantage to delaying such a fork for as long as possible by contributing to a major distro.
You don't seem to have understood the suggestion properly.
Consider a Firefox bugfix - every Ubuntu machine that is being used as a desktop will pop up an "update available" notice and allow the user to download the update. If the Ubuntu people count the number of downloads of *that update*, they'll get a reasonably accurate count of internet-connected Ubuntu desktops.
Unless you're downloading very large files, that's because of the latency rather than the bandwidth. Latency is horrible on *all* current cellular data standards, but it's not that much more horrible on the "2.5G" networks compared to the "3G" networks.
For those of you who really think that the difference between 2.5G and 3G is a deal breaker, I'd like to point something out:
Unless you are going to hook the phone up to your computer and use it as a modem, the difference doesn't matter. You've got a 2" screen with a relatively low resolution - even crappy video streaming will run over 2.5G (poorly, but who watches video on their phone anyway?).
If you could do VOIP or something over 3G that might make it matter, but the latency for cellular internet access is so horrible that it's not worth it. On my 3G Sprint (PowerVision) phone, I've never seen my ping get lower than 500ms - and I've even written midlets to test it.
Maybe 4G will matter, but the difference between 2.5G and 3G is *nothing* relevant for any phone usage pattern I can come up with. Even MP3 downloads are fast enough on 2.5G.
The fact that the people in Sweden let it get that far makes me sad. It's blatantly obvious that once you give the authorities censorship power, they'll use it against groups they don't like. The question of "what are they censoring" shouldn't even enter into it - giving those tools to the state is a horrible mistake.
Using a phone to try to solve that sort of problem is correct only when the people you talk to on the phone have the authority to fix your problem. The minute you realize that they don't have sufficient authority, you need to switch to postal mail with signature confirmation.
A good old "Dear , you are billing me for a service I never actually requested. Here are the details: ... Cease and desist immediately." on paper that they had to sign for will almost always produce the intended results. If it doesn't, at least you've got documentation to show a laywer.
How is that nuts? The hardware is pretty useless if it doesn't run the software you want to use.
No. In theory, as in practice, hardware only works when the OS supports it.
Windows doesn't support every piece of hardware that works with Linux. Why would you expect Linux to support every last piece of hardware that Windows does?
In that case, the hardware you bought to run Windows may not run any other OS. Sucks that you didn't plan ahead, but there's no way it's the fault of some other OS that it doesn't run on a machine designed specifically to run Windows and nothing else.
Perhaps, in the future, you'll consider software flexibility as one of your purchasing criteria.
You're right that C# and Java require super long columns to be even vaguely legible.
Personally, I'd say that's a flaw in the language. Java style interface naming, where the stock mutable string type is called "StringBuffer" (12 characters) is absurd, as is the idea that absolutely everything deserves a level of indentation (you can have "class { methood { lock { try {" before you even get to code). I've worked in a lot of programming languages, and none of them are perfect, but Java-like languages waste horizontal screen space like it was worthless.
What you're saying is "some random idea that no-one's ever tried, much less successfully marketed is the one true answer". That's obviously utter hogwash. All of the desktop operating systems that have ever seen any adoption whatsoever have a certain amount of underlying structure that a power user or developer can get at relatively easily.
From this perspective, any modern OS (Windows Vista, Mac OS X, Ubuntu 7.04) are all fine. They hide enough that the average user can do what they want without worrying about the OS, and they reveal enough that the programmers can write software.
Adobe is an important player in the Linux on the Desktop story, but I don't think they're as significant as you make them out to be. They're software is way to expensive and blatantly unnecessary for most users, and both Sun and Microsoft have declared open season on their only really general-user product (flash).
You asked the person you replied to to back up his claims - please back up the claims I quoted here. Actually, don't bother - it's impossible, because the only thing that can possibly back up "Linux is nich and will always be niche..." is a gut feeling.
This could be actually true, or it could be functionally true as the result of very conservative IT policies. Many industries have conservative IT policies, and for those industries change will come only very, very slowly.
But... here's the thing: Some specific systems in a conservative company's IT department have very little to do with overall market trends. Your use of Windows to track insurance will have precisely the same impact on Linux adoption that the mainframes had on Windows adoption in the early 90's - not much.
What's really important though is the fact that the computer industry isn't a single market. It's a whole bunch of smaller markets that all influence each-other to various degrees. And Linux is a solid, mainstream player everywhere from mainframes to cellphones. Feel free to bet that that won't translate into desktop gains, but that'd be a bet you'd lose.
This is very simple: BUY SUPPORTED HARDWARE.
Your expectation that an OS will work on random hardware makes no sense. You wouldn't expect Windows to run on a PPC Mac - why would you expect Ubuntu to give you 3D acceleration out of the box with an ATI Radeon X1300 (that has "Windows XP or higher" as a system requirement)?
When you buy a computer, you buy it for a specific purpose. You select hardware that can run the software you intend to use. The fact that some distribution of GNU/Linux doesn't provide perfect hardware support for the machine you bought to run Microsoft Windows shouldn't surprise you at all - in fact, every time you *succeed* at repurposing a machine in a manner that you didn't consider at the time of initial purchase you should be pleasantly surprised, since - in the great scheme of computer compatibility - that almost never happens.
Non-enthusiasts doing OS installs is a non-starter anyway. If your time is worth anything, you buy machines pre-installed - in which case any unsupported hardware means you return the computer as "broken". Dell sells Ubuntu desktops, as well as SLED and RHEL workstations.
Again, the "Consumer Desktop Market" has NOTHING to do with people self-installing operating systems, much less self-installing an operating systems on systems that weren't specifically built to run that OS.
You mean like a VIA C7?
Sure, we won't have the performance of today's 130 watt monsters in an 8 watt chip for a while, but you can sure get the performance of a slow Pentium IV in only a couple of watts.
Mips and Mflops are not useful benchmarks at all on modern processors. That measurement would be completely unaffected by any of the design decisions that effect actual code - stuff like cache and IO bandwidth just wouldn't matter at all. To use a bad car analogy, that's like asking what the sum of the theoretical maximum energy output per cycle of the cylinders - sure, it's a performance related metric, but not a useful one.
Routing through a protocal-specific smart cache sometimes does though.
Sure, the service needs to be offered at a finite number of IP addresses and the ISP can pretend to be a user to find out what they are, but even that method can be countered by silly techniques like user-partioning (different users see a different subset of servers based on a hash of their IP - so the ISP will only see 1/n of all of the servers). In the end, the best case for the ISP is that it turns into an arms race that eats endless man hours.
That's false. Unless the ISP wants to move to destination address whitelisting (i.e. you can only access approved servers), there's no way to reliably prevent people from using encrypted VPN tunnels for their P2P traffic. If set up properly, such tunnels can actually *improve* P2P download speeds in some cases.
Sorry... "obsolete" wasn't the right word. I was looking for "superseded by better deals".
The best price/performance at the moment is the AMD 3600. From there, the price performance ratios get progressively worse as performance increases. Other processors that aren't blatantly obsolete include: The AMD 5000, Intel 4300, AMD 5600, and a bunch of Intel processors costing over $200.
Yea, the national debit is about $5 trillion, so it would make a notifiable difference. The real question is this: What sort of difference would it make?
I suggest reading this article: http://en.wikipedia.org/wiki/Us_national_debt, and maybe watching this video: http://video.google.com/videoplay?docid=-905047436 2583451279. My personal point is simply that the situation is more complicated than "debt is always bad, no debt would be good".
The question isn't if the format will still be in use, the question is this: How hard will it be to write a converter to whatever the standard is in the future? With ODF, you need to uncompress a zip archive and parse some reasonably simple XML. With OOXML, you need to deal with all sorts of crap - and even if you have the standard, you won't necessarily be able to implement parts of it.