The idea is not just a formal API, but a formal API that doesn't change for a long (VERY VERY long if it is to be usefull at all) time. The problem is that you will be stuck with the mistakes for the same time.
You know, to address the concerns of the developers who want a binary interface for their drivers, you'd have to maintain driver API consistency for the duration of an entire major version number for the stable kernel. That's not exactly an eternity, or a huge burden.
Obviously, I need source code so vendors can't extort more money from me after the sale, or force me to load upgrades in order to obtain security fixes.
Ironically, that's exactly what not having a binary driver interface does to me, whenever I want to patch my kernel for security issues.
How about an OS where you have a choice? If you like binary drivers which may cause incompatibilities, instability, etc. then by all means, load 'em up! Otherwise, opt out of it and don't use them. Simple, huh?
Actually, it's just fucking retarded. Binary drivers don't cause incompatibilities just because they are binary drivers except on Linux. You aren't going to convince anyone, including naive or new users, that "binary drivers on our system are especially crappy!" is a good feature or something.
As someone who supports free software, and has struggled with NVIDIA's video drivers (and they're at least trying to meet us halfway by making it as easy as possible to install their closed-source driver under the current system) I can see the negative consequences of encouraging binary-only drivers.
Well, you sure misunderstood the situation.
The "negative consequences" you struggled with are due to Nvidia having to hack around Linux's lack of a binary interface for drivers. Nvidia's stuff installs fine on Windows and I don't hear a lot of complaints from my OS X-using friends.
Having to update your drivers when you update your kernel is a stupid, Linux-imposed necessity that just doesn't exist on most other modern OSes. It's not a shining victory for open source.
Nvidia has stuff from other companies in their drivers, so they are not allowed to release the source code.
Many companies believe that if they release the hardware specs then other vendors will be able to copy them cheaper. Of course, those other vendors can already reverse-engineer the hardware directly (I have a friend that does this for a living) so I don't know how much of an argument this really is.
In the case of ATI and Nvidia that is not really the issue. ATI relies on the DRI architecture, which sucks. If Nvidia were to release their GLX as open source ATI would adopt it and Nvidia would have given up a huge competitive advantage. It is in their commercial interest for ATI's drivers to continue sucking.
I don't buy the "stuff from other vendors" thing, Nvidia could buy SGI outright with their pocket change. The original Nvidia Linux drivers were a cooperative effort between Nvidia, SGI, and VA Linux. Nvidia doesn't have a lot of expensive commitments left in those drivers.
Personally I think they're right too: the high-handed approach may be more painful for some people right now, in the short-term, but in the long-term, it's going to lead to more open hardware.
We've already lived the long-term, and it hasn't happened that much. Anything that leads to increasingly open hardware specifications is going to involve a change, since we've been at a steady-state for a while now with the current approach.
Note that Windows pulls the same sort of API changing crap that Linux does.
No, they don't, unless you are putting Windows XP and Windows 95 in the same category. Just about any Windows 2000 driver will work on any version through service pack 4, but with Linux I need to recompile my drivers whenever a minor version number changes. Joy.
Dumping unwanted carbon dioxide seems like a perfect application for the first space elevator. It'd probably be a much easier experiment than an elevator built to carry a load. It just needs to be a very strong and simple tube, with less hardware tied into each end than the final elevator design will use.
I don't know what kind of chemical process you'd want to use to separate out the CO2 gas before blowing it up the tube, or exactly how much energy the whole thing would take. I'll bet that it will be a lot cheaper than liquefying and then moving the equivalent amount of CO2, once the elevator exists and its development costs are taken out of the equation.
Check out EROS [eros-os.org] for an implementation that exists now. Granted, EROS itself is no longer being developed, it was definitely around before this OS, and EROS has spawned some new projects (look on the link for links).
OMFG MS Sux they're ripping them off OMFG they've never done anything original OMG LOL!!!!!!11111111
Ok, so I'm dusted. I see that the most liberal of parties opposes what is effectively Free Speech and the party which brought us the Patriot Act is advocating the it.
Ah, don't worry. Both parties hold you in the same high esteem, and will put you behind barbed wire if you go to protest at one of their conventions.
Linux has a huge independent development community and more huge companies than it is easy to count behind it, and nobody can keep up with the pace of development. The GPL is a very important factor. It's the only partnership that would keep it fair for the big guys and the little ones at the same time.
Oddly enough, Bruce, the BSD and Apache communities are full of "little guy" developers who do not feel they are being treated unfairly.
It has to do with being allowed to mess up without the fear of having your every move dissected for weeks at a time by pundits, whose only motive is to use those mistakes and embarrass you.
If I say something "racist/imperialistic/unconstitutional" as the OP mentioned I should be embarassed. Especially if I'm dumb enough to say it in front of a camera.
If we expect our policiticans to start web-logging their daily thoughts, we're going to have to be a lot less hard on them about what they say. Our politicians, like the rest of the human race, are going to have ideas that, when fully thought out, are really bad. In maintaining weblogs some of these bad ideas are going to see international publication.
Three points. First, if someone has a truly abominable idea, call them on it. You don't owe them your "understanding". If they are in a position of authority you'd fucking well better set them straight.
Second, this isn't a particularly new challenge for a politician. You've got a lot more opportunity to edit yourself in a blog than you do on the Sunday talk shows, and with the comments turned off, you don't have anyone interacting with you. The only reason Hastert, for example, might be more stupidly off-the-cuff in his blog than on television would be that his handlers aren't coaching him for hours ahead of time. Poor guy.
Third, what a sad state we're in when we elect people so fucking dumb we feel like we have to go easy on them when they say something stupid. I guess that's not exactly a new phenomenon either, the press goes terribly easy on Bush, to the point of tweaking his quotes to make them sound less illiterate. But now the public is expected to do it too? Count me out.
The bit in the article about how XML will solve all our data interchange problems is particularly curious. C'mon, it's just text files with a bunch of angle brackets, when it gets right down to it.
Hey! Those text files have feelings, you insensitive clod.
First of all, this is a private high school. The First Amendment does not apply to private organizations, and even more so to religious private organizations *. Nor should it have to. If there is a problem with free speech, they can go to some other, possibly public school.
If they pull this nonsense, hopefully all the money they receive from the federal government will go to some other, possibly public school.
Obviously, this is not forgery with an intent to fool, but like posting unaltered dollar bill photographs on a website, it's at least uncool and asking for trouble.
Perhaps... of course, even ignoring what I assume you mean by "buttfuckathon", it's still much easier to make binary drivers for Windows since x86 is essentially the only architecture.
The CPU architecture isn't the issue at all. Having to rebuild the driver every time the kernel changes, or build some kind of hacky not-so-reliable wedge like Nvidia does, is the issue.
The manufacturer should be happy to do whatever the customers want
In principle I agree, although of course we'd need to pay for it. I'd fear a little if Nvidia (for example) calculated the cost to themselves of making their excellent OpenGL implementation public and then passed the cost on to us, the consumers.
apparently not enough other customers feel that way.
I have a feeling anything but some kind of electromagnetic rail thing is going to look pretty silly when you are seriously talking about achieving escape velocity.
Slashdot Mirror
You know, to address the concerns of the developers who want a binary interface for their drivers, you'd have to maintain driver API consistency for the duration of an entire major version number for the stable kernel. That's not exactly an eternity, or a huge burden.
Ironically, that's exactly what not having a binary driver interface does to me, whenever I want to patch my kernel for security issues.
Actually, it's just fucking retarded. Binary drivers don't cause incompatibilities just because they are binary drivers except on Linux. You aren't going to convince anyone, including naive or new users, that "binary drivers on our system are especially crappy!" is a good feature or something.
Well, you sure misunderstood the situation.
The "negative consequences" you struggled with are due to Nvidia having to hack around Linux's lack of a binary interface for drivers. Nvidia's stuff installs fine on Windows and I don't hear a lot of complaints from my OS X-using friends.
Having to update your drivers when you update your kernel is a stupid, Linux-imposed necessity that just doesn't exist on most other modern OSes. It's not a shining victory for open source.
In the case of ATI and Nvidia that is not really the issue. ATI relies on the DRI architecture, which sucks. If Nvidia were to release their GLX as open source ATI would adopt it and Nvidia would have given up a huge competitive advantage. It is in their commercial interest for ATI's drivers to continue sucking.
I don't buy the "stuff from other vendors" thing, Nvidia could buy SGI outright with their pocket change. The original Nvidia Linux drivers were a cooperative effort between Nvidia, SGI, and VA Linux. Nvidia doesn't have a lot of expensive commitments left in those drivers.
We've already lived the long-term, and it hasn't happened that much. Anything that leads to increasingly open hardware specifications is going to involve a change, since we've been at a steady-state for a while now with the current approach.
No, they don't, unless you are putting Windows XP and Windows 95 in the same category. Just about any Windows 2000 driver will work on any version through service pack 4, but with Linux I need to recompile my drivers whenever a minor version number changes. Joy.
Or better yet, gets random kiddies on the interweb to make it for free.
Dumping unwanted carbon dioxide seems like a perfect application for the first space elevator. It'd probably be a much easier experiment than an elevator built to carry a load. It just needs to be a very strong and simple tube, with less hardware tied into each end than the final elevator design will use.
I don't know what kind of chemical process you'd want to use to separate out the CO2 gas before blowing it up the tube, or exactly how much energy the whole thing would take. I'll bet that it will be a lot cheaper than liquefying and then moving the equivalent amount of CO2, once the elevator exists and its development costs are taken out of the equation.
Well, I found the trice, but it's pretty small. I don't think my freedom is going to fit.
If you spent any time talking to BMW salespeople you might.
I'm guessing those salespeople used to cop that attitude before Lexus became the benchmark for build quality...
Measured how? What were the bottlenecks?
OMFG MS Sux they're ripping them off OMFG they've never done anything original OMG LOL!!!!!!11111111
Too bad for SGI, a moral debt from Linux users and a pocketful of change will maybe get you a cup of coffee.
Ah, don't worry. Both parties hold you in the same high esteem, and will put you behind barbed wire if you go to protest at one of their conventions.
Oddly enough, Bruce, the BSD and Apache communities are full of "little guy" developers who do not feel they are being treated unfairly.
If I say something "racist/imperialistic/unconstitutional" as the OP mentioned I should be embarassed. Especially if I'm dumb enough to say it in front of a camera.
Three points. First, if someone has a truly abominable idea, call them on it. You don't owe them your "understanding". If they are in a position of authority you'd fucking well better set them straight.
Second, this isn't a particularly new challenge for a politician. You've got a lot more opportunity to edit yourself in a blog than you do on the Sunday talk shows, and with the comments turned off, you don't have anyone interacting with you. The only reason Hastert, for example, might be more stupidly off-the-cuff in his blog than on television would be that his handlers aren't coaching him for hours ahead of time. Poor guy.
Third, what a sad state we're in when we elect people so fucking dumb we feel like we have to go easy on them when they say something stupid. I guess that's not exactly a new phenomenon either, the press goes terribly easy on Bush, to the point of tweaking his quotes to make them sound less illiterate. But now the public is expected to do it too? Count me out.
Hey! Those text files have feelings, you insensitive clod.
If they pull this nonsense, hopefully all the money they receive from the federal government will go to some other, possibly public school.
Come on, there are plenty of counterexamples to that. You don't have to look too hard.
What a weiner you are.
How many rhetorical questions can you ask in one message? No, really? I'm asking?
The CPU architecture isn't the issue at all. Having to rebuild the driver every time the kernel changes, or build some kind of hacky not-so-reliable wedge like Nvidia does, is the issue.
In principle I agree, although of course we'd need to pay for it. I'd fear a little if Nvidia (for example) calculated the cost to themselves of making their excellent OpenGL implementation public and then passed the cost on to us, the consumers.
Good point.
I have a feeling anything but some kind of electromagnetic rail thing is going to look pretty silly when you are seriously talking about achieving escape velocity.