A small minority of people who can get their heads around the fact that the world's oil and gas reserves will not be here come 2100, and that it will thus take roughly the span of two intellectual generations to change our entire (think about this for a second) way of life or at least (yes, sceptics) that part of out lives which is centered around fossil energy use, but how do these two differ? If I remember correctly, the industrial revolution only really started picking up steam in the 1800s. So it only took 200 years for us to go from horses and buggies and blacksmiths, to automobiles and bullet trains. Technology is going to turn our lives upside down in the next century-- oil or no oil.
what worries me most is the notion that each country can solve its problems for itsel. People, the biggest challenge mankind faces in the 21st century is not "when and how will technology aid us" but rather, "will the transition be smooth, or will large groups of (poor)people miss out?". In one sense, countries can solve their energy problems for themselves. For example, Brazil has more or less moved all of their vehicles to locally produced ethanol. (They grow sugar cane and distill it.) Europe has a good train infrastructure that conserves energy.
In another sense, nobody has found a way to solve the global problems that we have: global warming and environmental degredation.
on a closing note i must say it seriously offends me that a story on North Korea's nuke draws approximately 3-4 times as many comments as something as paramount as this. But perhaps that largely betrays our inclination towards the short-term, the very mechanism that got us in this mess in the first place.;) Perhaps its because peak oil and other "futurist" topics have been discussion points for about two decades, whereas the korean nuke test is actual news?
You are right that the peak oil people are nuts. They fail to understand the ways that the economy can adapt to high oil prices.
Unfortunately, none of the alternatives you named are really all that desirable. Coal to gas conversion is very environmentally destructive-- and, of course, it contributes even more to global warming. All of the heavy oil products require even more energy to refine than light oil, which translates into massive inefficiency. Unfortunately, that is the future.
Ethanol-powered vehicles don't really reduce US oil consumption because US agriculture is massively dependent on petroleum-based fertizilers and other chemicals. Brazil, on the other hand, has a different climate which allows it to grow a lot of sugar cane, and there ethanol does help.
So basically, the environment is screwed, but the economy is not. Yay.
9 cents/kwh is just a back-of-the-envelope calculation.
Obviously, you could get more formal, and start considering the projected rate of inflation, bank interest rates, and returns on the stock market. People can, and do, consider these things when they're taking out mortgages on $100,000 homes. But for a $9,000 wind turbine, that seems like overkill.
I'd like to see exactly what savings account gives you 7% interest a year. I'm also curious why you decided to ignore taxes and inflation.
Nobody, however, knows just what those changes will be, or even which direction. To some extent you can mitigate that on the futures market. China has somewhere around 3 times the population of the US. India has a pretty hefty amount too, although I don't remember the exact number offhand. India especially will be a problem in the future because of geometric population growth-- they have little to no birth control. Both of these countries are industrializing. And so are a lot of other places in the world. An informed person can make an educated guess that oil, gas, and coal will be much more expensive in the future.
Even today, oil companies are going after resources that would have been too expensive to go after in the past. Eventually, there will be some theoretical point where the energy spent getting the resources out of the ground and refining them will be more than the energy gained by doing so, and people will start using renewable energy. What the earth will look like at that point, I leave up to your imagination.
First of all, I'm not speaking for my alma mater... my opinions are my own only.
I don't attach a whole lot of weight to strategic goals or mission statements. Money is the thing that talks, and NASA ain't getting it. The DOD, the NIH, and to a lesser extent the NSF are getting it. Go to any major university and see how many research projects are funded by the former agencies. Then see how few are funded by NASA. It may open your eyes.
If you are affiliated with the agency, then I wish you and your mission only the best. But don't try to act like Bush has been some great benefactor to space exploration or to science in general. We've all heard the speeches, and we all know where his priorities are.
Interesting..you seem to have read the President's Vision for Space Exploration but somehow only grabbed onto the "kill shuttle" part and missed the whole "build vehicles to explore the solar system part"... or was that inconvenient to your anti-Bush rant?
That's because killing the shuttle is real, whereas the rest is just talk. Talk is cheap.
Whoa there, cowboy. I'll sell you all the non-standard junk you want, and the evil gubment can't do nothing about it.
Want some 234.56 volt, hexagonal power outlets for your home? Only $.50 each. Of course, you'll have to buy all your appliances from me, but that's a small price to pay for FIGHTING the MAN. And my prices are very REASONABLE.
You are correct that running at the least possible privilege level is a good idea. In fact, Theo van Raadt reworked SSH to operate that way by default, some time ago. It's called "privilege separation."
The real flaw that's being discussed in this thread is a flaw in the X security model, and it's quite fundamental. Windows has some similar problems with "shatter attacks."
The hack that you are discussing to implement a trusted login path has already been implemented... it's called the "Secure Keyboard" option on the main menu of "xterm." When turned on, all keyboard events are sent exclusively to the xterm window, so the password sniffing attack is not possible. Of course, almost nobody knows about this, so the hack is basically useless. Windows, of course, implements its own trusted login path with CTRL+ALT+DELETE. This is something Windows got right.
A very informative comment. I especially like your description of type systems as theorem provers.
I agree that it would be nice to see some simpler, cleaner languages, but I'm not sure if that's going to happen any time soon. For example... suppose the people who design model checkers for SML only have to do only X man-years of work, while the people who do model checkers for C# have to do 100X. So what-- the model checkers for C# are still going to be cheap and widely available, because there is a market for it. And since people have existing infrastructure dependent on C#, they don't want to switch unless there's a reason.
On the other hand, there is probably a way to integrate more model checking into the programming language itself, and make it more friendly to model checkers. Type systems are really just one example of this. Another example is assertions, such as the pre- and post- conditions that Djikstra tried so hard to popularize. Another example is data encapsulation in object oriented languages.
The US has the best healthcare in the world for those who can afford it. For those who can't, there is really no medical safety net at all, except for emergency room visits. Also, medicines are often more expensive for patent reasons-- some of which are justified since they lead to new development.
It's hard to come up with a really good socialized healthcare system. The government is not really all that good at allocating resources, and often socialized medicine leads to doctor shortages, substandard care, etc.
I find it very hard to imagine where having source code that you can't run is helpful at all. Use your imagination. The Windows source code has helped a lot of crackers find exploits. The BSD TCP stack has been very helpful to people implementing their own TCP stacks. Source code is very useful as a reference implementation, for those who know how to read it.
Linus said it himself. He doesn't care about Linux's market share, he doesn't care if it displaces things. You said it yourself: you don't care if embedded linux lives or dies. And you don't understand Linus' point of view. Given that, it's pretty disingenious for you to paraphrase him here. HURD is that way --- >
There will likely be GPL-3 for other tools -- gcc, glibc, and so on. So just how useful is a kernel, by itself, in the embedded market? And suppose Linux was GPL-3 -- is Linux that much more useful than BSD there? What license gcc has is irrelevant because it will never be shipped on production boxes. glibc is under control of Ulrich Drepper, at Red Hat, and will probably stay with GPLv2, because doing otherwise would inconvenience too many important people. If not, there's always projects like uclibc waiting to take over.
These kinds of "minefields" were similar to ones being moaned about and then ignored when GPLv2 was chosen for the kernel. You're just scaremongering, and anyone bothering to go to that link you gave will find in the commentary a ton of criticism of the claims of Linus et al. It's not as one-sided as you hope it is. One license succeeded, so all licenses must. Good argument there.
GPLv3 never aimed to roll back DRM, just to regain the freedoms that the GPL used to give before the likes of Tivo, which uses one kind of DRM to scuttle those freedoms. They are just looking after their license. You clearly also know nothing of their beliefs about the service-based internet economy. In the recent GPLv3 meeting in Bangalore, they specifically pointed out that GPLv3 does not itself close any "ASP loophole" nor does it force it open. So what are their beliefs about the service-based economy? Or is that going to be decided by another session of the politburo?
I hope that business investment is scared away from any more Tivoization. Businesses know that they can Tivo-ize as much as they want with GPLv2 code or BSD code. And that's where they'll be putting their money and effort in the future. So the projects under the more liberal licenses will get bug reports and developer time, and the ones under GPLv3 will get neither. As it should be.
I suppose you think Tivoization is OK? I do, actually. That's because I know that I can always buy the hardware that I need to create my own Tivo, if the need arises. As soon as I can't buy that hardware, then we have a problem. And it's a problem that the GPLv3 can't do anything about.
And because so many will in fact not buy your kind of FUD, there remains the only true thing you have said: the fragmentation of the open source community; even more will be outraged about the concessions to Tivoization and believe more in the philosophy of free software. The GPLv3 is completely ineffective at stopping DRM, divisive to the community, and harmful to business.
COBOL was designed around the flawed idea that programming languages should be similar to natural languages. This, of course, makes about as much sense as the idea that computer mice should be similar to the kind of mice that nibble on cheese.
I agree that you need some kind of fixed-point representation for dollars and cents. In any civilized language, you would simply create a class that handles fixed-point math, and use that wherever you need fixed-point math. There isn't really any need for BCD-- it's just an encoding, and not a very efficient one, either.
I think you need to understand that people can write obfuscated code in any language. Code is readable because it was designed well, not because it was written in language $FOO. That being said, some languages are better than others. Generally a strong type system, terseness, and clearly defined semantics make a good language. As far as I know, COBOL doesn't have any of those. Sorry.
Oddly enough, America does a darn good job of automotive manufacturing and engineering. Honda and Toyota are expanding their North American operations because they get good quality work out of us folks here in Flyover, USA. The problem with American business is American government and American managers.
No, the big problem with US carmakers is how much they spend on health care and unionized labor. Back during the glory days of US industry, unions extracted some pretty significant concessions from Ford and GM. Basically, they agreed to extremely generous health care and pension plans for their workforce.
Now, their workforce is getting old rapidly, and the financial burden of all those old people is killing the companies, slowly. GM, for example, spends about $1,525 per vehicle produced on health care. Foreign companies are mostly free of these entanglements, and they can charge significantly less. Most countries have some kind of subsidized health care system, so companies can get away with not paying for workers' health care!
Slowly, the big US carmakers are forcing the unions to cave in on benefits, though. Basically they are giving an ultimatum: either find some way of reducing labor costs, or the company itself will go bankrupt, and the unions will get nothing at all. It's not a pretty picture.
I think the real reason why Honda and Toyota are expanding their US manufacturing and assembly operations is political. By throwing a few jobs our way, they keep the protectionists at bay. They can say, see, we ARE helping the local economy-- and make the loss of US dominance a less bitter pill to swallow. People aren't usually disposed to vote against the organization that employs them. But maybe this is just a conspiracy theory on my part.
I like your description of "the gated retirement community of open source." There are definitely some projects that seem to fade away. Sometimes this is because the developers move on, and sometimes progress just seems to grind to a halt.
Don't underestimate the inertia of established technologies, though. Some people are still using some variant of FORTRAN, the original compiled language developed in the 50s. Other organizations got stuck at the COBOL stage. There's still a small market for VAX gurus, VMS programmers-- you name it. No doubt academics will be the last people to abandon Java. I expect a lot of perl 5 code will keep chugging along for years-- maybe even decades, hidden here and there in server rooms around the world. Hell, bash scripting is still around, and that's about the worst language ever-- even BASIC might be better.
Computer enthusiasts tend to evaluate computer technologies in terms of how productive they are. But to corporate leaders, often the only really important number is the cost of switching. As long as that is high enough, it's better to keep patching and duct-taping the stuff you've got, rather than putting out the money for a new system. It might take years to see the benefits from a new system, and corporate leaders are often under pressure to show short-term results.
As long as there is enough money behind projects, they will NEVER die, no matter how boneheaded they may be. The next version of Windows will NEVER "enter Death Valley," because Microsoft needs to release something new, to get that nice load of cash from product upgrades, and to justify charging OAMs the Windows tax. Here on slashdot, we know that Windows 2000, XP, and Vista, are really just the same OS, give or take some candy-coated graphics and bloatware. But most consumers just buy the latest thing.
People in the know, and smart businesses, know that there's certain products and technologies that are just deprecated for new development work. For example, why use CVS, when Subversion is all that and the bag of chips. Why use PHP, when it's just a broken subset of Perl. And why use Excel macros when you have... anything else. But to pointy-haired bosses, the world is a lot less clear.
Note that BitKeeper's eventual "shenanigans" were predicted by RMS when everyone else called him paranoid for thinking a for-profit entity might someday change its mind. And RMS's shenanigans could have been predicted too. Frankly, I trust engineers and even businessmen more than I trust politicians.
In your contrived example, if a company goes to extraordinary lengths to prevent upgrades of ANY kind, then it's just like running GPLv3 code in ROM: that's ok. If they use measures specifically to prevent users from upgrading but allowing the company itself to, then that's NOT OK as per the draft of GPLv3. But now we are starting to make regulations based on people's intentions rather than on what they actually did. It is a slippery slope. And what if one part of the system, like the kernel, is unchangeable and in ROM, but the other is in RAM? There's an endless galaxy of hardware and software configurations that you can run software on. Why are we getting into the business of deciding what end users can do or not do?
The fact of the matter is this: media companies are pushing hard to get DRM and copy protection schemes implemented by hardware vendors. We in the open source community can fight that in a number of ways-- like creating cracks for the DRM, or lobbying politicians. But you can't fight it by changing the license of your software.
If they can't implement the security features that they need to, hardware companies will just drop GPLv3 software entirely. And that doesn't help anyone-- in fact it hurts the movement. It won't stop DRM, or the media companies. The only person it hurts is us, the end users and developers.
If the manufacturer doesn't like that, they are perfectly free to use any of the multitude of software available under other terms. Stop changing the subject. Of course I know that there is other software available. There's even Windows CE available. But most people would prefer that embedded Linux became the standard, not embedded Windows, and that just wouldn't happen if Linux accepted RMS's new regulations.
By Thomas Bushnell BSG: I have a lot of experience in dealing with RMS, and the very last think we want to do is get into a conversation about the question "Is GnuCash part of the GNU Project?" He will insist that the answer is yes, and begin imposing rules if he thinks it is necessary to force the "marriage". If the GnuCash developers protest hard enough, the result will only be that RMS will attempt to find a new gnucash maintainer (!) and appoint them the "official" GnuCash maintainer. Of course that doesn't affect anything directly, but it produces an instant code fork for no good reason.
The current situation is happily ambiguous, in which we are all (both GnuCash and the GNU Project) reaping the benefits of a marriage without needing to enter in to all the long-term promises which that entails...
Let's just make the website modifications because this will help us maintain peace. If RMS makes a truly unreasonable demand in the future, we can address that then. But the current request poses no real problems, and so in the interests of comity, I suggest we take it up, but without saying that we are doing so in response to his message.
> And Linus doesn't have a choice here. The kernel will likely always > be GPLv2, True.
> and if glibc and others switch to GPLv3, Linus won't be able to do a thing about it. Not true. There are already alternatives to glibc, like diet libc.
When BitKeeper tried to pull similar shenanigans, Linus dropped BitKeeper like a stone, and wrote his own source code management system, called "git." The same thing will happen to the GNU/zealotware, if necessary.
> > And you didn't answer the question. > It was a rhetorical question. No, it was a real and concrete question. Is a company acting in bad faith if it ships its embedded GPLv3 product with just exactly enough hard disk space to run the software it ships with, but no more? Or if the software is burned into EEPROM and unupgradable?
It's a question that you can't answer. Neither can Linus. And that's why we're worried.
> RMS is always the extreme. The GPLv3 is being drafted by more people than RMS. RMS is setting the direction, and making the final calls. Everyone else is just there to help point out self-contradictions and potential loopholes.
GCC came pretty near to forking around v2, with the egcs guys and all. It's hardly a shining example of "how to manage an open source project." That being said, I agree that it's a great achievement-- possibly even the most important GNU program.
GDB has always been a dog. It hardly works at all on multi-threaded programs, and still has major trouble with C++. Sometimes it's impossible to see the value of local variables. Other times the debugger itself segfaults. Not a happy situation. I know these things because I've had to use it at my company, and usually we end up using other debugging methods because it doesn't work!
EMACS might be a good program for some, but I don't use it, so I won't comment on it here.
In my experience, once someone calls someone a "republican," or "liberal," rational debate ends, and "football team" debate begins, where people howl at each other about "us versus them."
The anti-DRM provisions of the GPLv3 ARE a legal minefield. And Linus is not backing down, either. Those position papers represent a line drawn in the sand. He's deliberately refusing to legitimize the GPLv3 process by participating.
And even if Linus was willing to bend over backwards for Richard Stallman, he couldn't convert the existing Linux codebase over to GPLv3. He would have to get permission from each and every copyright holder to relicense. Read the discussion on LKML.
> > Are they going to go out of their way, and increase the cost of the > > product, to give this capability? > Well, they have another choice: They could just use code that isn't GPL'd. There's always BSD.
So Linus should just let embedded Linux die because of your zealotry? Unfuckingbelievable.
And you didn't answer the question.
> > Is the company acting in bad faith if it ships its embedded Linux with > > just exactly enough hard disk space to run the software it ships with, but no more? > Depends on the situation. It could be considered bad design -- they won't be > able to provide any firmware updates of their own.
Again, you didn't answer the question. And that's exactly the point. Nobody can answer these questions, because the GPLv3 is a confusing, overly broad mass of crap, that rivals the worst proprietary software licenses.
> Linus said, "Really, I'm not out to destroy Microsoft. That will just > be a completely unintentional side effect." > I think that applies here. GPLv3 is not out to destroy DRM, but that may > be a side effect, if such DRM depends heavily on open source software.
Except that RMS really is out to destroy DRM. For the record, I agree that DRM is bad. I just don't think a software license can stop it.
Doesn't Torvalds owe something to Stallman? He owes the GPLv2 to Stallman.
Did Linus forget where he came from? Um, Finland? What does that have to do with anything? If you're implying the Stallman came up with the original idea for Linux, or anything like that, you couldn't be more wrong. In fact, Linux was a competitor to the stillborn GNU operating system, which was HURD.
Linus was driven to create Linux partly by his frustrations with Professor Andy Tannenbaum's Minix, and partly by his own desire to create. At no time were RMS, ESR, or goats involved.
Slashdot Mirror
A small minority of people who can get their heads around the fact that the world's oil and gas reserves will not be here come 2100, and that it will thus take roughly the span of two intellectual generations to change our entire (think about this for a second) way of life or at least (yes, sceptics) that part of out lives which is centered around fossil energy use, but how do these two differ?
;)
If I remember correctly, the industrial revolution only really started picking up steam in the 1800s. So it only took 200 years for us to go from horses and buggies and blacksmiths, to automobiles and bullet trains. Technology is going to turn our lives upside down in the next century-- oil or no oil.
what worries me most is the notion that each country can solve its problems for itsel. People, the biggest challenge mankind faces in the 21st century is not "when and how will technology aid us" but rather, "will the transition be smooth, or will large groups of (poor)people miss out?".
In one sense, countries can solve their energy problems for themselves. For example, Brazil has more or less moved all of their vehicles to locally produced ethanol. (They grow sugar cane and distill it.) Europe has a good train infrastructure that conserves energy.
In another sense, nobody has found a way to solve the global problems that we have: global warming and environmental degredation.
on a closing note i must say it seriously offends me that a story on North Korea's nuke draws approximately 3-4 times as many comments as something as paramount as this. But perhaps that largely betrays our inclination towards the short-term, the very mechanism that got us in this mess in the first place.
Perhaps its because peak oil and other "futurist" topics have been discussion points for about two decades, whereas the korean nuke test is actual news?
You are right that the peak oil people are nuts. They fail to understand the ways that the economy can adapt to high oil prices.
Unfortunately, none of the alternatives you named are really all that desirable. Coal to gas conversion is very environmentally destructive-- and, of course, it contributes even more to global warming. All of the heavy oil products require even more energy to refine than light oil, which translates into massive inefficiency. Unfortunately, that is the future.
Ethanol-powered vehicles don't really reduce US oil consumption because US agriculture is massively dependent on petroleum-based fertizilers and other chemicals. Brazil, on the other hand, has a different climate which allows it to grow a lot of sugar cane, and there ethanol does help.
So basically, the environment is screwed, but the economy is not. Yay.
9 cents/kwh is just a back-of-the-envelope calculation.
Obviously, you could get more formal, and start considering the projected rate of inflation, bank interest rates, and returns on the stock market.
People can, and do, consider these things when they're taking out mortgages on $100,000 homes.
But for a $9,000 wind turbine, that seems like overkill.
I'd like to see exactly what savings account gives you 7% interest a year.
I'm also curious why you decided to ignore taxes and inflation.
Nobody, however, knows just what those changes will be, or even which direction. To some extent you can mitigate that on the futures market.
China has somewhere around 3 times the population of the US. India has a pretty hefty amount too, although I don't remember the exact number offhand. India especially will be a problem in the future because of geometric population growth-- they have little to no birth control.
Both of these countries are industrializing. And so are a lot of other places in the world.
An informed person can make an educated guess that oil, gas, and coal will be much more expensive in the future.
Even today, oil companies are going after resources that would have been too expensive to go after in the past. Eventually, there will be some theoretical point where the energy spent getting the resources out of the ground and refining them will be more than the energy gained by doing so, and people will start using renewable energy. What the earth will look like at that point, I leave up to your imagination.
First of all, I'm not speaking for my alma mater... my opinions are my own only.
I don't attach a whole lot of weight to strategic goals or mission statements.
Money is the thing that talks, and NASA ain't getting it. The DOD, the NIH, and to a lesser extent the NSF are getting it. Go to any major university and see how many research projects are funded by the former agencies. Then see how few are funded by NASA. It may open your eyes.
If you are affiliated with the agency, then I wish you and your mission only the best. But don't try to act like Bush has been some great benefactor to space exploration or to science in general. We've all heard the speeches, and we all know where his priorities are.
Interesting..you seem to have read the President's Vision for Space Exploration but somehow only grabbed onto the "kill shuttle" part and missed the whole "build vehicles to explore the solar system part"... or was that inconvenient to your anti-Bush rant?
That's because killing the shuttle is real, whereas the rest is just talk. Talk is cheap.
You are right that the unix administrator model is not very fine-grained... you are either root and all-powerful, or you are just a normal user.
But the real solution is not to make the current access control list system more complicated, but rather to move to a capabilities-based system.
None of this really has much to do with the problem we were talking about in X, which lets users read each others' keystrokes!
Whoa there, cowboy. I'll sell you all the non-standard junk you want, and the evil gubment can't do nothing about it.
Want some 234.56 volt, hexagonal power outlets for your home? Only $.50 each. Of course, you'll have to buy all your appliances from me, but that's a small price to pay for FIGHTING the MAN. And my prices are very REASONABLE.
You are correct that running at the least possible privilege level is a good idea. In fact, Theo van Raadt reworked SSH to operate that way by default, some time ago. It's called "privilege separation."
The real flaw that's being discussed in this thread is a flaw in the X security model, and it's quite fundamental. Windows has some similar problems with "shatter attacks."
The hack that you are discussing to implement a trusted login path has already been implemented... it's called the "Secure Keyboard" option on the main menu of "xterm." When turned on, all keyboard events are sent exclusively to the xterm window, so the password sniffing attack is not possible. Of course, almost nobody knows about this, so the hack is basically useless. Windows, of course, implements its own trusted login path with CTRL+ALT+DELETE. This is something Windows got right.
Ok, I have a program that, if it halts, causes a buffer overflow.
Otherwise, it doesn't.
Can your type system prove that there is no buffer overflow?
A very informative comment. I especially like your description of type systems as theorem provers.
I agree that it would be nice to see some simpler, cleaner languages, but I'm not sure if that's going to happen any time soon. For example... suppose the people who design model checkers for SML only have to do only X man-years of work, while the people who do model checkers for C# have to do 100X. So what-- the model checkers for C# are still going to be cheap and widely available, because there is a market for it. And since people have existing infrastructure dependent on C#, they don't want to switch unless there's a reason.
On the other hand, there is probably a way to integrate more model checking into the programming language itself, and make it more friendly to model checkers. Type systems are really just one example of this. Another example is assertions, such as the pre- and post- conditions that Djikstra tried so hard to popularize. Another example is data encapsulation in object oriented languages.
The US has the best healthcare in the world for those who can afford it. For those who can't, there is really no medical safety net at all, except for emergency room visits.
Also, medicines are often more expensive for patent reasons-- some of which are justified since they lead to new development.
It's hard to come up with a really good socialized healthcare system. The government is not really all that good at allocating resources, and often socialized medicine leads to doctor shortages, substandard care, etc.
I find it very hard to imagine where having source code that you can't run is helpful at all.
Use your imagination.
The Windows source code has helped a lot of crackers find exploits. The BSD TCP stack has been very helpful to people implementing their own TCP stacks.
Source code is very useful as a reference implementation, for those who know how to read it.
Linus said it himself. He doesn't care about Linux's market share, he doesn't care if it displaces things.
You said it yourself: you don't care if embedded linux lives or dies. And you don't understand Linus' point of view. Given that, it's pretty disingenious for you to paraphrase him here.
HURD is that way --- >
There will likely be GPL-3 for other tools -- gcc, glibc, and so on. So just how useful is a kernel, by itself, in the embedded market? And suppose Linux was GPL-3 -- is Linux that much more useful than BSD there?
What license gcc has is irrelevant because it will never be shipped on production boxes.
glibc is under control of Ulrich Drepper, at Red Hat, and will probably stay with GPLv2, because doing otherwise would inconvenience too many important people. If not, there's always projects like uclibc waiting to take over.
And a kernel by itself is actually very useful.
These kinds of "minefields" were similar to ones being moaned about and then ignored when GPLv2 was chosen for the kernel. You're just scaremongering, and anyone bothering to go to that link you gave will find in the commentary a ton of criticism of the claims of Linus et al. It's not as one-sided as you hope it is.
One license succeeded, so all licenses must. Good argument there.
GPLv3 never aimed to roll back DRM, just to regain the freedoms that the GPL used to give before the likes of Tivo, which uses one kind of DRM to scuttle those freedoms. They are just looking after their license. You clearly also know nothing of their beliefs about the service-based internet economy. In the recent GPLv3 meeting in Bangalore, they specifically pointed out that GPLv3 does not itself close any "ASP loophole" nor does it force it open.
So what are their beliefs about the service-based economy?
Or is that going to be decided by another session of the politburo?
I hope that business investment is scared away from any more Tivoization.
Businesses know that they can Tivo-ize as much as they want with GPLv2 code or BSD code. And that's where they'll be putting their money and effort in the future. So the projects under the more liberal licenses will get bug reports and developer time, and the ones under GPLv3 will get neither. As it should be.
I suppose you think Tivoization is OK?
I do, actually.
That's because I know that I can always buy the hardware that I need to create my own Tivo, if the need arises. As soon as I can't buy that hardware, then we have a problem. And it's a problem that the GPLv3 can't do anything about.
And because so many will in fact not buy your kind of FUD, there remains the only true thing you have said: the fragmentation of the open source community; even more will be outraged about the concessions to Tivoization and believe more in the philosophy of free software.
The GPLv3 is completely ineffective at stopping DRM, divisive to the community, and harmful to business.
Fair enough. GnuCash is a GNU project.
I still don't think very highly of RMS as a project manager.
COBOL was designed around the flawed idea that programming languages should be similar to natural languages. This, of course, makes about as much sense as the idea that computer mice should be similar to the kind of mice that nibble on cheese.
I agree that you need some kind of fixed-point representation for dollars and cents. In any civilized language, you would simply create a class that handles fixed-point math, and use that wherever you need fixed-point math. There isn't really any need for BCD-- it's just an encoding, and not a very efficient one, either.
I think you need to understand that people can write obfuscated code in any language. Code is readable because it was designed well, not because it was written in language $FOO. That being said, some languages are better than others. Generally a strong type system, terseness, and clearly defined semantics make a good language. As far as I know, COBOL doesn't have any of those. Sorry.
Oddly enough, America does a darn good job of automotive manufacturing and engineering. Honda and Toyota are expanding their North American operations because they get good quality work out of us folks here in Flyover, USA. The problem with American business is American government and American managers.
No, the big problem with US carmakers is how much they spend on health care and unionized labor. Back during the glory days of US industry, unions extracted some pretty significant concessions from Ford and GM. Basically, they agreed to extremely generous health care and pension plans for their workforce.
Now, their workforce is getting old rapidly, and the financial burden of all those old people is killing the companies, slowly. GM, for example, spends about $1,525 per vehicle produced on health care. Foreign companies are mostly free of these entanglements, and they can charge significantly less. Most countries have some kind of subsidized health care system, so companies can get away with not paying for workers' health care!
Slowly, the big US carmakers are forcing the unions to cave in on benefits, though. Basically they are giving an ultimatum: either find some way of reducing labor costs, or the company itself will go bankrupt, and the unions will get nothing at all. It's not a pretty picture.
I think the real reason why Honda and Toyota are expanding their US manufacturing and assembly operations is political. By throwing a few jobs our way, they keep the protectionists at bay. They can say, see, we ARE helping the local economy-- and make the loss of US dominance a less bitter pill to swallow. People aren't usually disposed to vote against the organization that employs them. But maybe this is just a conspiracy theory on my part.
I like your description of "the gated retirement community of open source." There are definitely some projects that seem to fade away. Sometimes this is because the developers move on, and sometimes progress just seems to grind to a halt.
Don't underestimate the inertia of established technologies, though. Some people are still using some variant of FORTRAN, the original compiled language developed in the 50s. Other organizations got stuck at the COBOL stage. There's still a small market for VAX gurus, VMS programmers-- you name it. No doubt academics will be the last people to abandon Java. I expect a lot of perl 5 code will keep chugging along for years-- maybe even decades, hidden here and there in server rooms around the world. Hell, bash scripting is still around, and that's about the worst language ever-- even BASIC might be better.
Computer enthusiasts tend to evaluate computer technologies in terms of how productive they are. But to corporate leaders, often the only really important number is the cost of switching. As long as that is high enough, it's better to keep patching and duct-taping the stuff you've got, rather than putting out the money for a new system. It might take years to see the benefits from a new system, and corporate leaders are often under pressure to show short-term results.
As long as there is enough money behind projects, they will NEVER die, no matter how boneheaded they may be. The next version of Windows will NEVER "enter Death Valley," because Microsoft needs to release something new, to get that nice load of cash from product upgrades, and to justify charging OAMs the Windows tax. Here on slashdot, we know that Windows 2000, XP, and Vista, are really just the same OS, give or take some candy-coated graphics and bloatware. But most consumers just buy the latest thing.
People in the know, and smart businesses, know that there's certain products and technologies that are just deprecated for new development work. For example, why use CVS, when Subversion is all that and the bag of chips. Why use PHP, when it's just a broken subset of Perl. And why use Excel macros when you have... anything else. But to pointy-haired bosses, the world is a lot less clear.
Yeah, PHP is strictly worse than Perl.
PHP is "training wheels without the bike."
Unfortunately, bad programming languages never really die. Even COBOL is still leaving its trail of slime at some mega-corporations.
Note that BitKeeper's eventual "shenanigans" were predicted by RMS when everyone else called him paranoid for thinking a for-profit entity might someday change its mind.
And RMS's shenanigans could have been predicted too.
Frankly, I trust engineers and even businessmen more than I trust politicians.
In your contrived example, if a company goes to extraordinary lengths to prevent upgrades of ANY kind, then it's just like running GPLv3 code in ROM: that's ok. If they use measures specifically to prevent users from upgrading but allowing the company itself to, then that's NOT OK as per the draft of GPLv3.
But now we are starting to make regulations based on people's intentions rather than on what they actually did. It is a slippery slope.
And what if one part of the system, like the kernel, is unchangeable and in ROM, but the other is in RAM? There's an endless galaxy of hardware and software configurations that you can run software on. Why are we getting into the business of deciding what end users can do or not do?
The fact of the matter is this: media companies are pushing hard to get DRM and copy protection schemes implemented by hardware vendors.
We in the open source community can fight that in a number of ways-- like creating cracks for the DRM, or lobbying politicians. But you can't fight it by changing the license of your software.
If they can't implement the security features that they need to, hardware companies will just drop GPLv3 software entirely. And that doesn't help anyone-- in fact it hurts the movement. It won't stop DRM, or the media companies. The only person it hurts is us, the end users and developers.
If the manufacturer doesn't like that, they are perfectly free to use any of the multitude of software available under other terms.
Stop changing the subject. Of course I know that there is other software available.
There's even Windows CE available. But most people would prefer that embedded Linux became the standard, not embedded Windows, and that just wouldn't happen if Linux accepted RMS's new regulations.
Read further.
By Thomas Bushnell BSG:
I have a lot of experience in dealing with RMS, and the very last
think we want to do is get into a conversation about the question "Is
GnuCash part of the GNU Project?" He will insist that the answer is
yes, and begin imposing rules if he thinks it is necessary to force
the "marriage". If the GnuCash developers protest hard enough, the
result will only be that RMS will attempt to find a new gnucash
maintainer (!) and appoint them the "official" GnuCash maintainer. Of
course that doesn't affect anything directly, but it produces an
instant code fork for no good reason.
The current situation is happily ambiguous, in which we are all (both
GnuCash and the GNU Project) reaping the benefits of a marriage
without needing to enter in to all the long-term promises which that
entails...
Let's just make the website modifications because this will help us
maintain peace. If RMS makes a truly unreasonable demand in the
future, we can address that then. But the current request poses no
real problems, and so in the interests of comity, I suggest we take
it up, but without saying that we are doing so in response to his
message.
> And Linus doesn't have a choice here. The kernel will likely always
> be GPLv2,
True.
> and if glibc and others switch to GPLv3, Linus won't be able to do a thing about it.
Not true. There are already alternatives to glibc, like diet libc.
When BitKeeper tried to pull similar shenanigans, Linus dropped BitKeeper like a stone, and wrote his own source code management system, called "git." The same thing will happen to the GNU/zealotware, if necessary.
> > And you didn't answer the question.
> It was a rhetorical question.
No, it was a real and concrete question.
Is a company acting in bad faith if it ships its embedded GPLv3 product with just exactly enough hard disk space to run the software it ships with, but no more? Or if the software is burned into EEPROM and unupgradable?
It's a question that you can't answer. Neither can Linus. And that's why we're worried.
> RMS is always the extreme. The GPLv3 is being drafted by more people than RMS.
RMS is setting the direction, and making the final calls.
Everyone else is just there to help point out self-contradictions and potential loopholes.
GCC came pretty near to forking around v2, with the egcs guys and all. It's hardly a shining example of "how to manage an open source project."
That being said, I agree that it's a great achievement-- possibly even the most important GNU program.
GDB has always been a dog. It hardly works at all on multi-threaded programs, and still has major trouble with C++.
Sometimes it's impossible to see the value of local variables. Other times the debugger itself segfaults. Not a happy situation.
I know these things because I've had to use it at my company, and usually we end up using other debugging methods because it doesn't work!
EMACS might be a good program for some, but I don't use it, so I won't comment on it here.
In my experience, once someone calls someone a "republican," or "liberal," rational debate ends, and "football team" debate begins, where people howl at each other about "us versus them."
The anti-DRM provisions of the GPLv3 ARE a legal minefield. And Linus is not backing down, either. Those position papers represent a line drawn in the sand. He's deliberately refusing to legitimize the GPLv3 process by participating.
And even if Linus was willing to bend over backwards for Richard Stallman, he couldn't convert the existing Linux codebase over to GPLv3. He would have to get permission from each and every copyright holder to relicense. Read the discussion on LKML.
> > Are they going to go out of their way, and increase the cost of the
> > product, to give this capability?
> Well, they have another choice: They could just use code that isn't GPL'd. There's always BSD.
So Linus should just let embedded Linux die because of your zealotry?
Unfuckingbelievable.
And you didn't answer the question.
> > Is the company acting in bad faith if it ships its embedded Linux with
> > just exactly enough hard disk space to run the software it ships with, but no more?
> Depends on the situation. It could be considered bad design -- they won't be
> able to provide any firmware updates of their own.
Again, you didn't answer the question. And that's exactly the point. Nobody can answer these questions, because the GPLv3 is a confusing, overly broad mass of crap, that rivals the worst proprietary software licenses.
> Linus said, "Really, I'm not out to destroy Microsoft. That will just
> be a completely unintentional side effect."
> I think that applies here. GPLv3 is not out to destroy DRM, but that may
> be a side effect, if such DRM depends heavily on open source software.
Except that RMS really is out to destroy DRM.
For the record, I agree that DRM is bad. I just don't think a software license can stop it.
Doesn't Torvalds owe something to Stallman?
He owes the GPLv2 to Stallman.
Did Linus forget where he came from?
Um, Finland? What does that have to do with anything?
If you're implying the Stallman came up with the original idea for Linux, or anything like that, you couldn't be more wrong.
In fact, Linux was a competitor to the stillborn GNU operating system, which was HURD.
Linus was driven to create Linux partly by his frustrations with Professor Andy Tannenbaum's Minix, and partly by his own desire to create. At no time were RMS, ESR, or goats involved.