Sometime in 1998, I attended a talk Linus gave in Silicon Valley. It was, I gather, his usual thing: he just answered questions, he didn't really have a speech. I don't remember how the subject came up, exactly, but he started talking about his goals for Linux.
He said, approximately, "I'm going to take over the world." And the audience laughed. "You think I'm kidding, I know you do. But I'm not. I'm going to take over the world." And I will tell you this: he meant it. There was a sharp look in his eye; an uber-geek at the height of his technical power, knowing that what he was doing was going to change everything. He knew it sounded ridiculous, but he was being absolutely honest, not joking in the slightest. He intended to dominate everything with his operating system.
I believe that, at the time, he thought of "the world" as being "the desktop". That was the center of computing back then; servers were rare, and the Internet was pretty young, but everyone had desktops. Linux was never really intended as anything except a desktop OS, at least in the beginning. But that early code was so amazingly reliable, compared to what Microsoft was offering, and so incredibly cheap, compared to what Sun was offering, that it ended up pressed into server duty. It was the accidental server; it fell sideways into that role, and ended up being one of the best solutions on offer, simply by the virtue of a clean design and good code.
So, here we are, fifteen years later. And, you know what? Linus *did* take over the world in most respects. But he didn't do it how I believe he expected to; I'm pretty sure he thought he would break the Windows monopoly directly, and take over everyone's home computers. But Microsoft has largely managed to defend itself there. Rather, Linux turned into an ecosystem, and it went around Microsoft, growing into many other markets. You still don't see it on the desktop, but it is absolutely ubiquitous everywhere else. Chances are quite good that any reasonably prosperous household in the First World is running Linux somewhere, possibly in multiple devices. They may even be running more Linux than they are Windows, without even realizing it. And if you use the Internet, you're talking, at least somewhat, with Linux machines. It's in phones, it's in tablets, it's in routers, it's in servers, it's in supercomputers. It's everywhere except the desktop.
Honestly, I think it would have continued to make inroads even there, but the GNOME team and Canonical went nuts chasing tablets, burning their existing users and giving them a horrible desktop product, trying for imaginary tablet customers that never materialized in any significant way. Ubuntu Linux, in 2010, was an outstanding desktop OS, and three years later, it still hasn't regained that usability, discoverability, and just general functionality again. If those desktop teams hadn't lost their collective minds, I think Linux would be doing well, even in the center of Microsoft's power.
That, however, didn't happen. None of the assaults on the Microsoft desktop fortress have ever been successful. Microsoft still rules there. But Linux is either a major player or completely dominant in every other computing market that's developed in the last fifteen years. And the advent of cheap-as-potato-chips computers, like the Raspberry Pi, will only increase that effect, as new markets arise, and Linux is adapted to fit them.
So, no, Linux has not won in the way that Linus originally intended. That battle was lost, decisively. But he and the kernel devs have thoroughly won the larger war.
Followup a couple of weeks later, for later readers.
I was at Fry's this weekend and I looked at their DLP sets. Every single one of them now looks fine to me. I can't see flicker at all. The first generation of DLP sets was horrid, but the ones that are out now are beautiful. I have very refresh-sensitive eyes, and these things are now rock-solid. I've always been the pickiest person I know about refresh, so if it's okay for me, I think it'll be okay for anyone.
As long as it's a recent model, I wouldn't be too worried about ordering blind... although you may prefer inspecting them up close. I found that the scaling of SD content tended to be rather ugly on some sets, so you may want to see for yourself. Overall, the picture quality is amazing, and I'd buy any of those sets in a heartbeat.
Honestly, I didn't even realize it had an ATI chipset to start with... somehow I missed that in the specs when I was ordering. I just assumed it was Intel. Stupid me.
Doesn't change the fact that that chipset was completely unstable, and in the SFF, ended up wasting a chassis, a tiny motherboard, and a power supply.
They should have fallen all over themselves to make it right, and instead left everyone twisting in the wind with useless hardware. ATI is on my shitlist for a LONG time.
Last year, I was trying to build a HTPC, and bought a little SFF box. It had an ATI chipset inside... what a complete piece of shit it was. The drivers were awful, and the USB never really worked right. The system was connected via USB wireless, and I could rarely copy more than a 25mb file or so before the ENTIRE USB SUBSYSTEM would lock up... wiping out keyboard, mouse, everything. Ended up having to do a hard power off every single time. Turns out this was a widely known problem and, to my knowledge, it was never fixed. That SFF was a complete waste of money, a total loss. I should have just lit a few hundred-dollar bills on fire.... at least it wouldn't have taken all the troubleshooting time.
After my previous experience with the dismal ATI graphic drivers, particularly in OpenGL, they are on my shitlist for at least the next three or four years. The hardware may be good, but who can tell with drivers that suck that badly?
I'd suggest steering WAY clear of any ATI chipset.
That's definitely not the entire explanation. When the Amiga showed an interlaced image, it was extremely visible... and not just to me, to everyone. (it was so bad that there was a cottage industry of 'flicker-fixing' devices.)
TV's don't show their interlace strongly, so I think the AC's suggestion that it's slower phosphors makes more sense.
That may very well be true, since I don't see flicker on TVs, but I certainly do on monitors. So you're probably right.
Fixed-frequency monitors, however, are really dangerous to use with X... I damaged a monitor that way, and eventually had to replace it a few months later. I was poor at the time and it was very painful.
I don't know how common my flicker sensitivity is, but since 85hz works very well (and 120hz is fantastic), and most monitors will go that fast, that's probably a better solution than fixed frequency. You're much less likely to melt the monitor from a weird X modeline.
Judging purely from anecdotal evidence (ie, just my eyes).... if you are sensitive to screen refresh rates on a CRT, avoid DLPs like the plague.
I've always been sensitive to monitor refresh rates.... I see flicker all the time. A regular 60hz refresh bothers me a lot. Gives me a headache. 75hz is the absolute minimum acceptable refresh on a CRT, and 85hz is fine, even under fluorescents. At home, I liked to shoot for 120Hz, because that matched well with nearly everything... it gave me the smoothest possible motion. I imagine that might not work under fluorescents, but I never had a work CRT that could go that fast. (I'm all LCD these days.) (Strangely, I have never been sensitive to the 30hz video refresh rate, and I have no idea why.)
DLP makes me want to claw out my eyes and run shrieking from the room. I can point out exactly which sets are DLP from a hundred feet away. If you are at all sensitive to CRT refresh, you MUST go see DLPs in person, and you absolutely must make sure you have an ironclad return policy. The saturation and color on DLPs is a little better than LCDs, and they tend to be cheaper, but a display that gives you motion sickness is no good, no matter how cheap it is.:-)
If, for some reason, you can't demo a set, then LCD is the safe choice... it will always work, and all your guests will be able to use it comfortably. Plasma is also a good choice, as long as you realize that it does wear out eventually. And, of course, there's always CRT-based units. They don't get as large as the other technologies, but they have amazing image quality and are very cheap, because they're the redheaded stepchild.... people think CRT is automatically inferior, just because it's old tech.
The major downside, at least to the Sony CRTs, is that they are incredibly heavy. You'll need help installing even a small screen. But the colors are rich and vibrant, the blacks are dead black, and the resolution is far better than the CRTs of old.
They're misrepresenting the product. I have to wonder if they were paid for this review.
Of course they were.... If not outright in cash, by getting early access to hot new stuff.
It's also not unknown for a manufacturer to 'accidentally' forget to ask for their stuff back if you write a really glowing review of it.
NVidia appears to have shills working the forums, hired via some marketing agency. This is a hard thing to prove conclusively, but there was at least one documented case a couple months back, so assuming that there are more seems reasonable. This could be just a new evolution in that process... shill websites.
It's interesting that ALL of the preview articles I've read have involved massively overclocking the Conroe, and then breathing hard about the OCed numbers. ALL of them. Nobody seems to talk about stock speeds much at all.
I've seen quite a number of complaints about gcc optimizations... I don't know if this is a recent thing or not, but apparently gcc is prone to 'optimize' code in ways that causes it to run faster, but give the wrong result, at least some of the time.
Many of the Open Source projects are extremely robust and reliable. Samba, Apache, PostgreSQL, MySQL, Postfix... all examples of very high quality software, at least as good as most commercial software, if not better. The basic system utilities are often overlooked when thinking about open source reliability, but packages like vim and bash and perl are all extremely powerful and incredibly robust. Most stuff that's actually gotten to the point of being a defacto part of Unix is so reliable that you could literally bet your life on it. (Well, you could bet your life on Windows too, but you'd probably lose rather more often than you'd like: once.)
That said, however, there are a number of packages that are less robust. PHP has been a mass of security bugs forever. Ethereal is constantly in the news for similar reasons. The Linux kernel has been much less reliable in the 2.6 series. (which really sucks, because no matter how good your other utilities are, if the kernel's not up, they don't run.) Those are just a few examples, ones that I've personally had to deal with... I'm sure there are many, many others.
Being open source tends to make products better, but it's no panacea. Good testing is important. A bad development process will result in poor products, even if the individual code pieces are brilliant. Over time, the groups that pay attention to process and focus on quality will stand out. That's how, I believe, Apache took over the world.
Reliability isn't a fundamental quality of either open OR closed source. It's a combination of factors: the skill of the coders and the quality of the development/testing process are probably the two main ones. This fellow is trying to paint the open software coders as amateurs -- which they often are -- and implies that only software you pay for can be reliable. Both types CAN BE reliable or unreliable. But open code is a strong advantage, to the point that the less-disciplined free software teams can often do a better job than the highly-structured closed-source ones.
That said, many open source projects aren't run very well, and it's hard to tell whether software is stable based purely on IRC chats and web forums. Fortunately, there's a nice answer to that: the distros. This is where organizations like RedHat and SUSE shine. They try to cut through the BS, find the stuff that really works, and ship that. Whether or not they succeed in that endeavor is up to the market to decide. Is it willing to pay them for their time and expertise in assembling their distros?
By and large, they seem to be doing pretty well. All revenue these companies earn is entirely voluntary on the part of the customer, since their products are free software. The fact that so many people do pay is a strong sign that they're doing a good job.
Maybe it's not as common as it used to be... at one time, everyone in the US promised that. I wasn't aware that it had really changed.... but then again, I didn't read my current provider's TOS carefully either. (shame on me, eh?)
As far as I know, Speakeasy still sells 'unlimited' plans, explicitly unlimited. It's your bandwidth, you paid for it. Their AUP is pretty much, "don't spam, and don't do illegal stuff." They're pricey, but they're REALLY good.
Pacbell was also like that when they first started selling DSL... they treated it just like their larger corporate customers. But I think that changed later on. I don't know what they're like now, but when they first got into the low-cost, high-speed DSL business, they were extremely good... incredibly low latency, because they were all ATM on the backbone. I was like four hops, and 8 or 10ms, from MAE West. That meant something, at the time.:-)
Their AUP was pretty similar to Speakeasy's, but they probably went to the nastier 'no servers allowed' type contracts later, once they figured out that they weren't dealing with professionals anymore. I have no direct evidence of that, but it IS Pacbell... once the DSL business got big enough to get on management's radar, they'd have screwed it all up.
Well, one of us has a very large misunderstanding of the NVidia kernel binary blob, and I'm honestly not sure who it is.
The way *I* learned it was that the binary blob was the Windows core driver, and the glue layer provided entry points that the Linux kernel expected.
If that's the case, then Linus can say whatever the hell he wants, it's still not a derived work of the Linux kernel. If it were, NVidia couldn't distribute it without source, now could they?
Since NVidia can legally distribute their code without falling under the GPL, I would be quite confident in my own ability to do so as well. If it's a derived work, they can't distribute it. So it's not a derived work. If it's not a derived work, then I can do whatever the hell I want with it, as long as the NVidia license allows redistribution. (I haven't checked that.)
Even if I distribute it as a binary kernel module in/lib/modules somewhere, it's still not a derived work of Linux. If it is, then NVidia's got a big, big problem.
I should amend that banks can still outright FAIL, if they're run poorly... the Fed doesn't protect against that. What the Fed provides is, essentially, unlimited cash to cover withdrawals if a bank experiences a run. This costs the bank money, they don't provide the service for free.. I think it's called the 'overnight funds rate', but I'm not entirely positive.
Yes, actually, that's precisely what they will do. With the Fed backing them, there's no such thing as a run on a bank anymore. NO bank in the Fed system will ever 'not have enough money'.... they'll print/lend as much as they need.
You can't print bandwidth, so it's really not a very good analogy anyway.:)
What the providers really fear is that people will actually start using what they've been told they already have.
They've got giant pipes running into everyone's houses, and business models predicated on the fact that most people don't use them. So they tell everyone 'unlimited bandwidth!' when in fact they cannot provide this.
The tiered-internet thing is just a way to punish the people who actually use the bandwidth they were already sold. And an attempt to enact a tax on those who dare to actually provide data that's interesting enough that lots of their customers want it, all at the same time.
In the case of NVidia's driver, this does not apply.
The part that is a derived work of Linux is freely available under GPL. It is wrapper code that interfaces with a driver for Windows. That binary blob is probably a derived work of Windows, but in no way is it a derived work of Linux.
Distributing a Linux kernel with code, the Linux interface module with code, and the NVidia binary module without code fully satisfies the GPL, because the NVidia binary blob is not derivative of any GPL code. Period.
By your argument, the LinuxAnt people, who get wireless drivers working by the same method (an opensource wrapper on Windows drivers), are also illegal because they're not distributing the source code for the Windows drivers. That's a rather unique and novel theory, as far as I can see.
If I understand correctly, someone with some technical chops can remotely point an antenna at your Bluetooth device in use, force a reintroduction (which would probably cause a brief interruption of service), and by eavesdropping that introduction, snoop on all subsequent traffic. It's apparently not even that hard to do.
I suggest, therefore, not typing critical passwords on Bluetooth keyboards, and not saying incriminating things on Bluetooth headsets.
That's with standard antennas. If you deploy an antenna with enough gain, and a loud enough signal, you could talk to a Bluetooth device from kilometers away.
Kilometers would be extremely expensive, but a few hundred meters should be doable with Pringles-can-level equipment.
Security people don't like it because it's totally insecure.
Awhile ago, they'd figured out how to hijack the 'introduction' process between two devices, but it only worked during introduction, so it wasn't a terribly useful attack.
But then someone else discovered how to force a reintroduction at will. Between the two holes, Bluetooth has no security whatsoever, unless something pretty dramatic has changed in the last six months.
Be sure to read replies, as I've not been paying close attention to this. But as far as I know... don't send anything over Bluetooth you care about.
Furthermore, I'm shocked that Slashdot had the courage to post something that wasn't only making fun of the Wii for it's name! Could it be that we're actually going to get to read about its performance and abilities instead of just griping about its poor name choice? That's outlandish!
Indeed. I, for one, think we should stop teasing reporters about waving their Wiis around.
Slashdot Mirror
OK, I'm going to delurk for this one.
Sometime in 1998, I attended a talk Linus gave in Silicon Valley. It was, I gather, his usual thing: he just answered questions, he didn't really have a speech. I don't remember how the subject came up, exactly, but he started talking about his goals for Linux.
He said, approximately, "I'm going to take over the world." And the audience laughed. "You think I'm kidding, I know you do. But I'm not. I'm going to take over the world." And I will tell you this: he meant it. There was a sharp look in his eye; an uber-geek at the height of his technical power, knowing that what he was doing was going to change everything. He knew it sounded ridiculous, but he was being absolutely honest, not joking in the slightest. He intended to dominate everything with his operating system.
I believe that, at the time, he thought of "the world" as being "the desktop". That was the center of computing back then; servers were rare, and the Internet was pretty young, but everyone had desktops. Linux was never really intended as anything except a desktop OS, at least in the beginning. But that early code was so amazingly reliable, compared to what Microsoft was offering, and so incredibly cheap, compared to what Sun was offering, that it ended up pressed into server duty. It was the accidental server; it fell sideways into that role, and ended up being one of the best solutions on offer, simply by the virtue of a clean design and good code.
So, here we are, fifteen years later. And, you know what? Linus *did* take over the world in most respects. But he didn't do it how I believe he expected to; I'm pretty sure he thought he would break the Windows monopoly directly, and take over everyone's home computers. But Microsoft has largely managed to defend itself there. Rather, Linux turned into an ecosystem, and it went around Microsoft, growing into many other markets. You still don't see it on the desktop, but it is absolutely ubiquitous everywhere else. Chances are quite good that any reasonably prosperous household in the First World is running Linux somewhere, possibly in multiple devices. They may even be running more Linux than they are Windows, without even realizing it. And if you use the Internet, you're talking, at least somewhat, with Linux machines. It's in phones, it's in tablets, it's in routers, it's in servers, it's in supercomputers. It's everywhere except the desktop.
Honestly, I think it would have continued to make inroads even there, but the GNOME team and Canonical went nuts chasing tablets, burning their existing users and giving them a horrible desktop product, trying for imaginary tablet customers that never materialized in any significant way. Ubuntu Linux, in 2010, was an outstanding desktop OS, and three years later, it still hasn't regained that usability, discoverability, and just general functionality again. If those desktop teams hadn't lost their collective minds, I think Linux would be doing well, even in the center of Microsoft's power.
That, however, didn't happen. None of the assaults on the Microsoft desktop fortress have ever been successful. Microsoft still rules there. But Linux is either a major player or completely dominant in every other computing market that's developed in the last fifteen years. And the advent of cheap-as-potato-chips computers, like the Raspberry Pi, will only increase that effect, as new markets arise, and Linux is adapted to fit them.
So, no, Linux has not won in the way that Linus originally intended. That battle was lost, decisively. But he and the kernel devs have thoroughly won the larger war.
I DID screw up... I did 6 times as wide instead of 5. 216 is correct for 6, but it's still the wrong answer. :-)
The actual interesting part would be that if the composition were the same, a spherical object six times as wide would have 216 times as much mass.
In other words, if it hit at anywhere close to the same speed, this one was A LOT more destructive.
Followup a couple of weeks later, for later readers.
I was at Fry's this weekend and I looked at their DLP sets. Every single one of them now looks fine to me. I can't see flicker at all. The first generation of DLP sets was horrid, but the ones that are out now are beautiful. I have very refresh-sensitive eyes, and these things are now rock-solid. I've always been the pickiest person I know about refresh, so if it's okay for me, I think it'll be okay for anyone.
As long as it's a recent model, I wouldn't be too worried about ordering blind... although you may prefer inspecting them up close. I found that the scaling of SD content tended to be rather ugly on some sets, so you may want to see for yourself. Overall, the picture quality is amazing, and I'd buy any of those sets in a heartbeat.
If at first you don't succeed, redefine success.
Honestly, I didn't even realize it had an ATI chipset to start with... somehow I missed that in the specs when I was ordering. I just assumed it was Intel. Stupid me.
Doesn't change the fact that that chipset was completely unstable, and in the SFF, ended up wasting a chassis, a tiny motherboard, and a power supply.
They should have fallen all over themselves to make it right, and instead left everyone twisting in the wind with useless hardware. ATI is on my shitlist for a LONG time.
Last year, I was trying to build a HTPC, and bought a little SFF box. It had an ATI chipset inside... what a complete piece of shit it was. The drivers were awful, and the USB never really worked right. The system was connected via USB wireless, and I could rarely copy more than a 25mb file or so before the ENTIRE USB SUBSYSTEM would lock up... wiping out keyboard, mouse, everything. Ended up having to do a hard power off every single time. Turns out this was a widely known problem and, to my knowledge, it was never fixed. That SFF was a complete waste of money, a total loss. I should have just lit a few hundred-dollar bills on fire.... at least it wouldn't have taken all the troubleshooting time.
After my previous experience with the dismal ATI graphic drivers, particularly in OpenGL, they are on my shitlist for at least the next three or four years. The hardware may be good, but who can tell with drivers that suck that badly?
I'd suggest steering WAY clear of any ATI chipset.
That's definitely not the entire explanation. When the Amiga showed an interlaced image, it was extremely visible... and not just to me, to everyone. (it was so bad that there was a cottage industry of 'flicker-fixing' devices.)
TV's don't show their interlace strongly, so I think the AC's suggestion that it's slower phosphors makes more sense.
That may very well be true, since I don't see flicker on TVs, but I certainly do on monitors. So you're probably right.
Fixed-frequency monitors, however, are really dangerous to use with X... I damaged a monitor that way, and eventually had to replace it a few months later. I was poor at the time and it was very painful.
I don't know how common my flicker sensitivity is, but since 85hz works very well (and 120hz is fantastic), and most monitors will go that fast, that's probably a better solution than fixed frequency. You're much less likely to melt the monitor from a weird X modeline.
Judging purely from anecdotal evidence (ie, just my eyes).... if you are sensitive to screen refresh rates on a CRT, avoid DLPs like the plague.
:-)
I've always been sensitive to monitor refresh rates.... I see flicker all the time. A regular 60hz refresh bothers me a lot. Gives me a headache. 75hz is the absolute minimum acceptable refresh on a CRT, and 85hz is fine, even under fluorescents. At home, I liked to shoot for 120Hz, because that matched well with nearly everything... it gave me the smoothest possible motion. I imagine that might not work under fluorescents, but I never had a work CRT that could go that fast. (I'm all LCD these days.) (Strangely, I have never been sensitive to the 30hz video refresh rate, and I have no idea why.)
DLP makes me want to claw out my eyes and run shrieking from the room. I can point out exactly which sets are DLP from a hundred feet away. If you are at all sensitive to CRT refresh, you MUST go see DLPs in person, and you absolutely must make sure you have an ironclad return policy. The saturation and color on DLPs is a little better than LCDs, and they tend to be cheaper, but a display that gives you motion sickness is no good, no matter how cheap it is.
If, for some reason, you can't demo a set, then LCD is the safe choice... it will always work, and all your guests will be able to use it comfortably. Plasma is also a good choice, as long as you realize that it does wear out eventually. And, of course, there's always CRT-based units. They don't get as large as the other technologies, but they have amazing image quality and are very cheap, because they're the redheaded stepchild.... people think CRT is automatically inferior, just because it's old tech.
The major downside, at least to the Sony CRTs, is that they are incredibly heavy. You'll need help installing even a small screen. But the colors are rich and vibrant, the blacks are dead black, and the resolution is far better than the CRTs of old.
They're misrepresenting the product. I have to wonder if they were paid for this review.
Of course they were.... If not outright in cash, by getting early access to hot new stuff.
It's also not unknown for a manufacturer to 'accidentally' forget to ask for their stuff back if you write a really glowing review of it.
NVidia appears to have shills working the forums, hired via some marketing agency. This is a hard thing to prove conclusively, but there was at least one documented case a couple months back, so assuming that there are more seems reasonable. This could be just a new evolution in that process... shill websites.
It's interesting that ALL of the preview articles I've read have involved massively overclocking the Conroe, and then breathing hard about the OCed numbers. ALL of them. Nobody seems to talk about stock speeds much at all.
I've seen quite a number of complaints about gcc optimizations... I don't know if this is a recent thing or not, but apparently gcc is prone to 'optimize' code in ways that causes it to run faster, but give the wrong result, at least some of the time.
Many of the Open Source projects are extremely robust and reliable. Samba, Apache, PostgreSQL, MySQL, Postfix... all examples of very high quality software, at least as good as most commercial software, if not better. The basic system utilities are often overlooked when thinking about open source reliability, but packages like vim and bash and perl are all extremely powerful and incredibly robust. Most stuff that's actually gotten to the point of being a defacto part of Unix is so reliable that you could literally bet your life on it. (Well, you could bet your life on Windows too, but you'd probably lose rather more often than you'd like: once.)
That said, however, there are a number of packages that are less robust. PHP has been a mass of security bugs forever. Ethereal is constantly in the news for similar reasons. The Linux kernel has been much less reliable in the 2.6 series. (which really sucks, because no matter how good your other utilities are, if the kernel's not up, they don't run.) Those are just a few examples, ones that I've personally had to deal with... I'm sure there are many, many others.
Being open source tends to make products better, but it's no panacea. Good testing is important. A bad development process will result in poor products, even if the individual code pieces are brilliant. Over time, the groups that pay attention to process and focus on quality will stand out. That's how, I believe, Apache took over the world.
Reliability isn't a fundamental quality of either open OR closed source. It's a combination of factors: the skill of the coders and the quality of the development/testing process are probably the two main ones. This fellow is trying to paint the open software coders as amateurs -- which they often are -- and implies that only software you pay for can be reliable. Both types CAN BE reliable or unreliable. But open code is a strong advantage, to the point that the less-disciplined free software teams can often do a better job than the highly-structured closed-source ones.
That said, many open source projects aren't run very well, and it's hard to tell whether software is stable based purely on IRC chats and web forums. Fortunately, there's a nice answer to that: the distros. This is where organizations like RedHat and SUSE shine. They try to cut through the BS, find the stuff that really works, and ship that. Whether or not they succeed in that endeavor is up to the market to decide. Is it willing to pay them for their time and expertise in assembling their distros?
By and large, they seem to be doing pretty well. All revenue these companies earn is entirely voluntary on the part of the customer, since their products are free software. The fact that so many people do pay is a strong sign that they're doing a good job.
If you don't get the subtext, you now are automatically guilty if you have encrypted files, and must prove yourself innocent.
Britain has been edging very close to police-state status. If this law passes, it will cross that line once and for all.
Maybe it's not as common as it used to be... at one time, everyone in the US promised that. I wasn't aware that it had really changed.... but then again, I didn't read my current provider's TOS carefully either. (shame on me, eh?)
:-)
As far as I know, Speakeasy still sells 'unlimited' plans, explicitly unlimited. It's your bandwidth, you paid for it. Their AUP is pretty much, "don't spam, and don't do illegal stuff." They're pricey, but they're REALLY good.
Pacbell was also like that when they first started selling DSL... they treated it just like their larger corporate customers. But I think that changed later on. I don't know what they're like now, but when they first got into the low-cost, high-speed DSL business, they were extremely good... incredibly low latency, because they were all ATM on the backbone. I was like four hops, and 8 or 10ms, from MAE West. That meant something, at the time.
Their AUP was pretty similar to Speakeasy's, but they probably went to the nastier 'no servers allowed' type contracts later, once they figured out that they weren't dealing with professionals anymore. I have no direct evidence of that, but it IS Pacbell... once the DSL business got big enough to get on management's radar, they'd have screwed it all up.
Well, one of us has a very large misunderstanding of the NVidia kernel binary blob, and I'm honestly not sure who it is.
/lib/modules somewhere, it's still not a derived work of Linux. If it is, then NVidia's got a big, big problem.
The way *I* learned it was that the binary blob was the Windows core driver, and the glue layer provided entry points that the Linux kernel expected.
If that's the case, then Linus can say whatever the hell he wants, it's still not a derived work of the Linux kernel. If it were, NVidia couldn't distribute it without source, now could they?
Since NVidia can legally distribute their code without falling under the GPL, I would be quite confident in my own ability to do so as well. If it's a derived work, they can't distribute it. So it's not a derived work. If it's not a derived work, then I can do whatever the hell I want with it, as long as the NVidia license allows redistribution. (I haven't checked that.)
Even if I distribute it as a binary kernel module in
I should amend that banks can still outright FAIL, if they're run poorly... the Fed doesn't protect against that. What the Fed provides is, essentially, unlimited cash to cover withdrawals if a bank experiences a run. This costs the bank money, they don't provide the service for free.. I think it's called the 'overnight funds rate', but I'm not entirely positive.
Yes, actually, that's precisely what they will do. With the Fed backing them, there's no such thing as a run on a bank anymore. NO bank in the Fed system will ever 'not have enough money'.... they'll print/lend as much as they need.
:)
You can't print bandwidth, so it's really not a very good analogy anyway.
This is preaching to the choir, but bits is bits.
What the providers really fear is that people will actually start using what they've been told they already have.
They've got giant pipes running into everyone's houses, and business models predicated on the fact that most people don't use them. So they tell everyone 'unlimited bandwidth!' when in fact they cannot provide this.
The tiered-internet thing is just a way to punish the people who actually use the bandwidth they were already sold. And an attempt to enact a tax on those who dare to actually provide data that's interesting enough that lots of their customers want it, all at the same time.
In the case of NVidia's driver, this does not apply.
The part that is a derived work of Linux is freely available under GPL. It is wrapper code that interfaces with a driver for Windows. That binary blob is probably a derived work of Windows, but in no way is it a derived work of Linux.
Distributing a Linux kernel with code, the Linux interface module with code, and the NVidia binary module without code fully satisfies the GPL, because the NVidia binary blob is not derivative of any GPL code. Period.
By your argument, the LinuxAnt people, who get wireless drivers working by the same method (an opensource wrapper on Windows drivers), are also illegal because they're not distributing the source code for the Windows drivers. That's a rather unique and novel theory, as far as I can see.
Ads served = money. That's why they do it that way.
If I understand correctly, someone with some technical chops can remotely point an antenna at your Bluetooth device in use, force a reintroduction (which would probably cause a brief interruption of service), and by eavesdropping that introduction, snoop on all subsequent traffic. It's apparently not even that hard to do.
I suggest, therefore, not typing critical passwords on Bluetooth keyboards, and not saying incriminating things on Bluetooth headsets.
That's with standard antennas. If you deploy an antenna with enough gain, and a loud enough signal, you could talk to a Bluetooth device from kilometers away.
Kilometers would be extremely expensive, but a few hundred meters should be doable with Pringles-can-level equipment.
Security people don't like it because it's totally insecure.
Awhile ago, they'd figured out how to hijack the 'introduction' process between two devices, but it only worked during introduction, so it wasn't a terribly useful attack.
But then someone else discovered how to force a reintroduction at will. Between the two holes, Bluetooth has no security whatsoever, unless something pretty dramatic has changed in the last six months.
Be sure to read replies, as I've not been paying close attention to this. But as far as I know... don't send anything over Bluetooth you care about.