Well, you're closer, but you're still not on the money.
It sounds like either Linux-ABI steps on SCO patents, or certain customers are shipping SCO libraries to run on top of Linux-ABI (which is outright copyright violation).
They don't *need* patents. They just need for the SCO C library licence to say that programs developed using it may only be run on genuine SCO Unix. In that case, anyone running such a program would be in trouble -- they wouldn't have to redistribute it.
I suspect the background for this story is that a few long-standing SCO customers with an eye to the future have had a bright young nerd look at how hard it would be to get their vertical application to run on Linux instead of SCO. (Perhaps it's a dental surgery management suite running on Ingres or something similar.) Probably in many cases the customer has a binary app without source access, but that can be fixed with Linux-ABI. It's probably not so hard in most cases.
It's a good deal for the customer: they cut out their SCO licence costs, they get a platform with a bright future, and they have much less trouble finding people who can support and enhance it.
This is a bit bad for SCO, though. Once word gets back to HQ that this is happening, they start to think about methods that can be used to keep their customers locked in. One technique is to exploit the licence that the customer's application vendor originally signed to get the SCO libraries. If SCO were smart enough to put in a "this can only be used on SCO" clause, then they're set!
Anybody who has the source for their applications should be easily able to move to Linux, and probably most of the commercial applications like Oracle already have native ports. Linux-ABI and this licensing strategy really just apply to people with legacy SCO apps who can't, or don't want, to port to Linux.
Microsoft could use such a clause in the Office (or DirectX or MSVC Runtime) licences to put an end to all this Wine, Crossover and Transgaming nonsense, if they wanted to. I think there are enough precedents for that kind of restriction in software licences that it would be possible. For example, lots of driver software comes with a licence saying it may only be used with the vendor's original software. I think this technique is a terrible abuse of customers, like most proprietary software licences. But it would probably work to shake down some more money.
I don't know how many people care about just seeing source code, under glass as it were. People and companies tend to be interested in free software either for the sake of freedom, or because it's free, or because they can take it and make something better. Being charged for look-don't-touch doesn't really satisfy any of those desires.
I suppose people wanting to do security audits might care, but really the number of organizations in the world with the budget to seriously audit Windows can be counted on your fingers. I think even most government bodies apart from the spooks wouldn't care much -- after all, they're all using Outlook now despite the known problems. Auditing isn't going to tell them anything new.
I suppose it might be helpful in debugging problems in interaction with Windows. You can imagine device vendors wanting more access than they have at the moment. But I suspect the NDAs will be pretty restricted. Debugging Samba interoperability is probably right out.
We're in the early stages of Microsoft's stumblings towards respecting the freedom of their customers. When democracy won out in eastern Europe the countries weren't destroyed outright, but rather they came around to a different way of working. (Imperfect analogy, but humor me.) Satisfying as it might be to imagine Microsoft bankrupt, a more likely optimistic outcome is that in a few years they'll be a semi-open-source company, along the lines of Sun or IBM.
Re:Four years and half too late.
on
Ark Linux
·
· Score: 2
What Linux needs is more announcements of distributions "focussed on ease of use for desktop users." There's no need to actually ship anything -- just making announcements seems to be sufficient to get a Slashdot story. And you can spend hours posting here, rather than staring blankly at ark.c in Pico!
Or, somehow make your site really be an important reference for that topic -- add lots of reference information about widgets, and encourage people to link to you.
It seems to me (IANAL) that the root problem is Australia's regressive defamation laws. The fact that the publication took place via the internet is only a (headline-catching) detail.
Truth, the public interest, or lack of malice are not always sufficient defences in Australian law, which is a terrible thing for free speech.
The result is that defamation law is often used by the rich and powerful to deter criticisms. It is seldom helpful to ordinary people whose reputations are attacked unfairly....
Law reform commissions have been advocating reform of defamation law for decades. Possible changes include:
* public figure defence so that it's possible to make stronger criticisms of those with more power;
* adjudication outside courts, to reduce court costs;
* elimination of monetary pay-outs, requiring instead apologies published of equal prominence to the original defamatory statements.
In spite of widespread support for reform among those familiar with the issues, Australian law remains much the same. That's because it serves those with the greatest power, especially politicians who make the law and groups that use it most often.
If things proceed in this direction I would not be surprised to see some business publications (Barrons, The Economist,...) banned by the publisher from distribution in Australia for fear of libel suits. They are already self-censored in countries such as Singapore and Thailand from time to time.
You've never been to an American high school science class, I take it.
No, I went to school in Australia, where the teachers did a good job of describing the scientific method in the way I said.
The whole system is built on and depends on the students not questioning the teachers.
Perhaps this is why America is far worse than every other western country in belief in the more ridiculous religious doctrines? I seem to remember a recent NAS survey finding that something like 40% of Americans believed in creation along the lines of the Genesis myth within the past 6000 years.
(Instead of "we all evolved randomly", a thought of "what would it take for intelligent design".)
Well, there is basically no credible evidence for "Intelligent Design". It's just a few kooks and a larger number of credulous followers. As I point out elsewhere, none of them have any kind of explanation of *how* it might have occurred. The National Academy of Sciences recently released a statement to this effect. So it's not bad to occasionally discuss wrong or unfounded theories in science classes, but there are so many that you can't cover all of them. I think my chemistry teacher talked about phlogiston, for example.
There is no religious bias in science. It proceeds merely by examining the evidence and comparing theories. If it should happen that people's previous beliefs are proved wrong, that's just tough.
Teaching superstition in schools is incredibly harmful to the development of rational thought. "Perhaps a god created the universe?" "Perhaps sickness is caused by the Evil Eye, or impure thoughts?" "Perhaps food poisoning occurs when it's prepared by a menstruating woman?" The problem with all these ideas is not merely that they are factually incorrect, but that teaching them encourages children to believe doctrine rather than to enquire after truth.
The rocks roll down the hill because of gravity as opposed to the will of angry-godling, but that doesn't mean that angry-godling could not have started them rolling.
Actually, yes, it does mean that the godling didn't start them rolling. Once we understand the naturalistic explanation about erosion, soil mechanics, catastrophe theory and so on, there is simply no need for the angry godling any more, and in the absence of other evidence sensible people delete him from their beliefs.
Oh, and as the Universe has no real center, we can place the center wherever it's most convenient for us.
A few years ago, people used biblical arguments to "prove" that the Earth was absolutely stationary, and the rest of the universe rotated around it. Indeed, the whole weight of the "infallible(tm)" Catholic church was on this doctrine. Of course now it's ridiculous.
Creationism is already seen as pretty similar by most educated people and eventually even the southern US will catch up.
god did it through means that we can't possibly understand
OK, perhaps you're not saying that the means are impossible to understand. However, until you explain in detail how you think a god did create life, it is not really fair for you to ridicule somebody else's attempt.
So, how is it god created life? Do you believe he fashioned a body out of clay, and "breathed life into it"? How did the clay turn into flesh?
Or did he create a simple cell, and then guide evolution? What was it that caused the atoms to move into that configuration? Did they appear from nothing? Or were they floating around, and grabbed by some kind of electromechanical force? By what means were the dynamic processes set into motion. (Was the chicken, or the egg created first?)
I seriously doubt if you can provide any answers to this that sound more credible than mainstream scientific explanations for the origin of life. But I would be interested to hear you try. (And I'm not being sarcastic.)
The issue of origins is a philosophical one, not a scientific one. If you assert that matter or the universe has existed forever, that is not scientifically falsifiable.
Religions have often tried to mark off particular questions as being "not in the proper domain of science", but they have consistently been proved wrong. Science is merely rational ordered enquiry. Any question about the universe is in principle open to such enquiry.
So for example we can say that "the universe is no more than 6000 years old", and attempt to falsify that theory by finding matter older than 6000 years. Proving it infinitely old is perhaps proving a negative, but we might demonstrate convincingly that it has apparently always been in a metastable state.
except Jesus Christ who demonstrated His infinite power by being raised from the dead.
I'm sorry, but that's mere assertion! Can't you see that people of other religions will hold equally strong beliefs in the existence of their own mythological beings? Many people honestly believe the Dalai Lama is a reincarnation of a supernatural being, who has exerted an enormous beneficial influence on the world. Aside from your personal belief, what makes your position any different?
Stories of torment, sacrifice, and resurrection are a dime a dozen amongst saviour religions. (Odin, for example.) Having founders of the religion swear to the truth is pretty unconvincing.
The heart of the scientific method is trying to open your own assumptions to criticism.
However, Christian believers can't do that, because they think their eternal life depends on solid belief in their doctrine. Without abandoning their faith, they are not able to genuinely consider whether it is true or not.
His followers overturned the most powerful government in the world without use of force
"Without use of force" is an exaggeration.
By a parallel argument, do you think that (largely nonviolent) Buddhist evangelism of Asia proves that their beliefs are universally true? If not, what is the difference?
(btw, this article will probably be archived soon, but if you're finding the debate as interesting as i am please mail me.)
Good science classes shouldn't say that "science" is an irrefutable, infalible body of facts. "Science" is a process of successive approximation whereby we gradually, haltingly proceed towards more reliable and general understandings of the law. (It's religions that arrogantly claim to have a privileged claim to universal truth: infallible popes, and so on.) A good science course ought to get around to discussing Popper and philosophy of science.
The backwards extrapolation of evolution to explain the origin of our species and all species is not proven science.
Nothing can be proven beyond any doubt, but evolution is pretty well established, as much as say the atomic theory of chemistry, or gravity. You may not want to accept it, but practically every unbiased (non-fundamentalist) person who examines the evidence will accept it.
the origin of life is one of those few parts where science cannot be tested against religion.
I can't see any prima facie reason why it can't be tested. The origin of life is just an event that occurred at some time in the past. As with any past even we can try to understand it by trying to find and interpret evidence of the events, or by trying to reproduce a similar situation, or indeed just by eliminating impossibilities, amongst other methods.
A similar argument could have been made a few years ago about thunder being caused by electrons vs angry gods, or about the geocentric universe. What makes your case any different?
It's not reasonable to compare the likelihood of "an intelligent designer" vs a hypothesis of evolution from (say) ferrite clays, because they're not at comparable levels of detail and specificity.
To take your analogy of a crime scene: the police may have some trouble working out exactly which human committed the crime, and how. But that doesn't mean that we need to start assuming that the crime was committed by mythological beings. The general hypothesis ("it was done by a human") is well-established, even if the details ("it was done by the butler") are not yet.
At a very broad level of detail, we can weigh up "there is an intelligent designer", vs "there is no intelligent design but merely process". Some arguments can be made each way, but on the whole it seems unnecessary to introduce mysterious invisible actors without overwhelming evidence.
If you want to get specific, then mainstream science can provide a number of possible explanations for how various processes occur: life arrived on asteroids, evolved in clay, evolved in tide pools, etc. If you want to fairly challenge them, you ought to provide a countertheory at a comparable level of detail. Don't just say "God did it", but rather answer: By what mechanism did god move atoms around to create life? By what means did he associate souls and animals? Where did god come from? Why is your idea of god any more convincing than any of the thousands of others?
To be taken seriously, you have to answer all those questions in a way that is scientifically falsifiable. Creationists tend to intentionally leave the details fairly vague and hard to refute, saying that god did it through means that we can't possibly understand. That may or may not be true, but since it's not open to criticism or evaluation there is not much point in considering it.
I posed the question because I was trying to work out why the previous poster said
Knuth's MMX uses self-modifying code to store the return address in procedure calls.
So I suppose either they misunderstood/misremembered Knuth, or there is some positive aspect to it that Knuth can see and I cannot.
Or perhaps they were talking about MIX (the ~1970s processor) not MMIX (the modern RISC one)! Yes, now that I think about it that sounds like a more likely explanation -- it was probably just a typo.
Self-modifying code for procedure returns might be much more representative of the old mainframe machines that Knuth was modelling in his original books. Given current programming languages and machine architectures (split I/D caches) it looks a bit perverse.
Of course, some versions of FORTRAN couldn't do recursive calls to a function, which would seem to be the problem with directly modifying the function's code to insert the return address.
I guess you could put a self-modifying trampoline on the stack containing the return address, but... why not just store the address then?
Suppose a spammer attaches a long list of neutral words to his e-mail in order to 'dilute' the bad words. This way some innocent words might get assigned positive spam probability thus resulting in false positives later.
No, that shouldn't help them.
Broadly, these filters look for "interesting" words: either ones that are often found in spam but not in nonspam, or vice versa.
The word "the" often occurs in both types; the word "slartibartfast" rarely in either. Therefore both of them are neutral.
The word "stop-on-solib-event" has never occurred in my spam yet, but if they started adding it (through a web robot) then eventually the filter would come to think that it was neutral, rather than a nonspam indicator as at present.
Interestingly, misspellings can be a really good indicator of spamminess. Many spams are sent repeatedly, either through multiple addresses, or because they're chain letters, or just because the stupid spammers send them repeatedly. If there's a characteristic misspelling that hasn't occurred anywhere else, it becomes a good way to identify the spam.
Another nice thing about bogofilter (and possibly others) is that it considers origin IPs and domains along with body words. So things like seed.net.tw are likely to be dumped unless they have some other strongly nonspammy words. For me, AOL and Yahoo have slightly spammy smells -- more because they're often forged than because much spam originates there, I think.
Eventually these filters may be defeated, but I think they will work well for a while.
I have a theory that the reason many people want to filter spam is not just because of bandwidth or time, but rather because the moronic writing and presentation is an insult to a thinking reader. It is quite literally junk mail. If by trying to get past these filters spammers have to act more like reasonable humans it's probably a good thing.
This is less of a blunt instrument than DNS blacklists and therefore probably a good thing.
Here's the point: if I develop confidence in bogofilter and build a good database, then I can use it as an ingress filter on some large (>8000 members) mailing lists that I run. Every spam that gets through there costs us bandwidth, and wastes the time of every subscriber. In addition, spam messages which do get through to the list typically generate additional traffic, either by people complaining about spam, or in bounces from recipients whose servers reject it.
So the return from a good spamstopper is potentially enormous.
Any messages which are false positives can be handled by some other means; e.g. forwarding to a postmaster for review.
You know, dihydrogen oxide is pretty dangerous in high concentrations too. The length of a chemical name is a poor indication of its toxicitity.
More seriously, although Halon contains halogens (chlorine, fluorine, and bromine), so do most of the plastics used in a server room. You don't want to be around when they're burning.
Incidentally, this is why there are separate "plenum rated" ethernet cables: they're made from chemicals that won't break down into anything *too* nasty when they burn, and therefore they can legally be used in air vents and similar spaces.
Many people seem to believe this myth that halon removes oxygen from the air, and that it is deadly to humans. It's simply not true.
A fire requires three things: fuel, oxygen, and heat. A fire can be stopped in four ways: by removing fuel (e.g. shutting off a gas tap), by removing oxygen (e.g. smothering in CO2 or a fire blanket), by cooling (e.g. with a water hose), or by interrupting the chemical reaction. Halon is an example of this last strategy.
If you wanted to exclude oxygen from the machine room, the most economical and effective approach would probably be to dump CO2 into it. (This is why most office fire extinguishers use CO2 -- it works nicely on small fires and is safe around electrical equipment.) The problem is that in a confined space this would also asphyxiate any humans. (As another example of a confined space containing humans, halon is used on airplanes.)
Halon gas is used in concentrations of about 5%, at which level it is not harmful to humans for short periods. It's roughly as poisonous as other short organic molecules. It could be compared to the stuff you smell when using spraycans or working on a motor vehicle: you shouldn't do it all day or intentionally inhale it, but brief infrequent exposure is not particularly harmful. (Presumably your machine room doesn't catch fire every day...)
This is how it works, as I understand it: halon molecules compete with other chemicals in the room to preferentially absorb energy and free radicals from ignition points. This interrupts the chain reaction, so that energy liberated by burning goes into decomposing halon, rather than into setting more stuff alight. For this reason it can work at low concentrations.
The products produced when halon burns are bad for you, but as another poster pointed out they're no worse than other chemicals produced in a machine room fire.
Production of new halon systems has been tightly restricted because of their potential to damage the ozone layer. That says nothing about whether they're harmful for humans. (Indeed, the high UV flux that causes ozone breakdown in combination with CFCs would be pretty bad for you...)
Regardless of the chance of conviction, perhaps Alan has better things to do than put himself through the stress, risk and expense of a criminal trial? The Sklyarov trial hardly inspires confidence in the US "justice" system.
Alan needs to turn over kernel development to someone with more balls than he has.
I don't see anyone stopping you from creating a fork... If you think your enormous testicles make you better qualified, please go ahead.
Yes, something like this is more general. distcc is only useful for compiling C and C++.
However the drawback is that it requires you to use a shared filesystem and to have the same compilers and headers installed on all machines. If you already have that situation --/home NFS-mounted, and a single software image -- then it's great.
However many systems are not so tightly controlled. distcc is much less intrusive to install -- you can distribute jobs to a machine administered by somebody else without requiring them to start using NFS or stop upgrading libraries.
Yeah, linking is now the dominant factor in the Kernel Compile Wars. A 64-way POWER4 machine running 2.5.x can build it in about 4 seconds last time I checked, and most of that is linking.
A paralellizable linker might make a nice Masters project for someone.
The fact that ClearCase seems to slow down file operations by a factor of about 2 to 10 makes this argument not exactly compelling. I've seen it take 40 seconds to list a short directory, and on very well-hung machine.
Perhaps there are good things about ClearCase, but speed certainly isn't one of them.
That said, the function you described is basically what ccache does -- only without requiring stupid kernel tricks.
Slashdot Mirror
According to ABC local radio, a fourth person has been found dead, apparently trapped in her house.
Wow, nice call.
It's amusing, and humbling, how often Slashdotters (or humains in general) manage to hold contradictory opinions like this.
--
John Paul Gaultier, pron Goat-ie-er. Concidence?
They don't *need* patents. They just need for the SCO C library licence to say that programs developed using it may only be run on genuine SCO Unix. In that case, anyone running such a program would be in trouble -- they wouldn't have to redistribute it.
I suspect the background for this story is that a few long-standing SCO customers with an eye to the future have had a bright young nerd look at how hard it would be to get their vertical application to run on Linux instead of SCO. (Perhaps it's a dental surgery management suite running on Ingres or something similar.) Probably in many cases the customer has a binary app without source access, but that can be fixed with Linux-ABI. It's probably not so hard in most cases.
It's a good deal for the customer: they cut out their SCO licence costs, they get a platform with a bright future, and they have much less trouble finding people who can support and enhance it.
This is a bit bad for SCO, though. Once word gets back to HQ that this is happening, they start to think about methods that can be used to keep their customers locked in. One technique is to exploit the licence that the customer's application vendor originally signed to get the SCO libraries. If SCO were smart enough to put in a "this can only be used on SCO" clause, then they're set!
Anybody who has the source for their applications should be easily able to move to Linux, and probably most of the commercial applications like Oracle already have native ports. Linux-ABI and this licensing strategy really just apply to people with legacy SCO apps who can't, or don't want, to port to Linux.
Microsoft could use such a clause in the Office (or DirectX or MSVC Runtime) licences to put an end to all this Wine, Crossover and Transgaming nonsense, if they wanted to. I think there are enough precedents for that kind of restriction in software licences that it would be possible. For example, lots of driver software comes with a licence saying it may only be used with the vendor's original software. I think this technique is a terrible abuse of customers, like most proprietary software licences. But it would probably work to shake down some more money.
After three months of Antarctic cold and darkness,.. they look pretty.
I don't know how many people care about just seeing source code, under glass as it were. People and companies tend to be interested in free software either for the sake of freedom, or because it's free, or because they can take it and make something better. Being charged for look-don't-touch doesn't really satisfy any of those desires.
I suppose people wanting to do security audits might care, but really the number of organizations in the world with the budget to seriously audit Windows can be counted on your fingers. I think even most government bodies apart from the spooks wouldn't care much -- after all, they're all using Outlook now despite the known problems. Auditing isn't going to tell them anything new.
I suppose it might be helpful in debugging problems in interaction with Windows. You can imagine device vendors wanting more access than they have at the moment. But I suspect the NDAs will be pretty restricted. Debugging Samba interoperability is probably right out.
We're in the early stages of Microsoft's stumblings towards respecting the freedom of their customers. When democracy won out in eastern Europe the countries weren't destroyed outright, but rather they came around to a different way of working. (Imperfect analogy, but humor me.) Satisfying as it might be to imagine Microsoft bankrupt, a more likely optimistic outcome is that in a few years they'll be a semi-open-source company, along the lines of Sun or IBM.
Obviously these folks need to read the Open Source HOWTO.
What Linux needs is more announcements of distributions "focussed on ease of use for desktop users." There's no need to actually ship anything -- just making announcements seems to be sufficient to get a Slashdot story. And you can spend hours posting here, rather than staring blankly at ark.c in Pico!
Or, somehow make your site really be an important reference for that topic -- add lots of reference information about widgets, and encourage people to link to you.
If you want to know how the folks at Google feel about Scientology, try a Google search for "goatse". No, really! Look at the suggested category...
:-)
Nice hack
Truth, the public interest, or lack of malice are not always sufficient defences in Australian law, which is a terrible thing for free speech.
Excellent summary of Australian defamation law:
If things proceed in this direction I would not be surprised to see some business publications (Barrons, The Economist,
You've never been to an American high school science class, I take it.
No, I went to school in Australia, where the teachers did a good job of describing the scientific method in the way I said.
The whole system is built on and depends on the students not questioning the teachers.
Perhaps this is why America is far worse than every other western country in belief in the more ridiculous religious doctrines? I seem to remember a recent NAS survey finding that something like 40% of Americans believed in creation along the lines of the Genesis myth within the past 6000 years.
(Instead of "we all evolved randomly", a thought of "what would it take for intelligent design".)
Well, there is basically no credible evidence for "Intelligent Design". It's just a few kooks and a larger number of credulous followers. As I point out elsewhere, none of them have any kind of explanation of *how* it might have occurred. The National Academy of Sciences recently released a statement to this effect. So it's not bad to occasionally discuss wrong or unfounded theories in science classes, but there are so many that you can't cover all of them. I think my chemistry teacher talked about phlogiston, for example.
There is no religious bias in science. It proceeds merely by examining the evidence and comparing theories. If it should happen that people's previous beliefs are proved wrong, that's just tough.
Teaching superstition in schools is incredibly harmful to the development of rational thought. "Perhaps a god created the universe?" "Perhaps sickness is caused by the Evil Eye, or impure thoughts?" "Perhaps food poisoning occurs when it's prepared by a menstruating woman?" The problem with all these ideas is not merely that they are factually incorrect, but that teaching them encourages children to believe doctrine rather than to enquire after truth.
The rocks roll down the hill because of gravity as opposed to the will of angry-godling, but that doesn't mean that angry-godling could not have started them rolling.
Actually, yes, it does mean that the godling didn't start them rolling. Once we understand the naturalistic explanation about erosion, soil mechanics, catastrophe theory and so on, there is simply no need for the angry godling any more, and in the absence of other evidence sensible people delete him from their beliefs.
Oh, and as the Universe has no real center, we can place the center wherever it's most convenient for us.
A few years ago, people used biblical arguments to "prove" that the Earth was absolutely stationary, and the rest of the universe rotated around it. Indeed, the whole weight of the "infallible(tm)" Catholic church was on this doctrine. Of course now it's ridiculous.
Creationism is already seen as pretty similar by most educated people and eventually even the southern US will catch up.
god did it through means that we can't possibly understand
OK, perhaps you're not saying that the means are impossible to understand. However, until you explain in detail how you think a god did create life, it is not really fair for you to ridicule somebody else's attempt.
So, how is it god created life? Do you believe he fashioned a body out of clay, and "breathed life into it"? How did the clay turn into flesh?
Or did he create a simple cell, and then guide evolution? What was it that caused the atoms to move into that configuration? Did they appear from nothing? Or were they floating around, and grabbed by some kind of electromechanical force? By what means were the dynamic processes set into motion. (Was the chicken, or the egg created first?)
I seriously doubt if you can provide any answers to this that sound more credible than mainstream scientific explanations for the origin of life. But I would be interested to hear you try. (And I'm not being sarcastic.)
The issue of origins is a philosophical one, not a scientific one. If you assert that matter or the universe has existed forever, that is not scientifically falsifiable.
Religions have often tried to mark off particular questions as being "not in the proper domain of science", but they have consistently been proved wrong. Science is merely rational ordered enquiry. Any question about the universe is in principle open to such enquiry.
So for example we can say that "the universe is no more than 6000 years old", and attempt to falsify that theory by finding matter older than 6000 years. Proving it infinitely old is perhaps proving a negative, but we might demonstrate convincingly that it has apparently always been in a metastable state.
except Jesus Christ who demonstrated His infinite power by being raised from the dead.
I'm sorry, but that's mere assertion! Can't you see that people of other religions will hold equally strong beliefs in the existence of their own mythological beings? Many people honestly believe the Dalai Lama is a reincarnation of a supernatural being, who has exerted an enormous beneficial influence on the world. Aside from your personal belief, what makes your position any different?
Stories of torment, sacrifice, and resurrection are a dime a dozen amongst saviour religions. (Odin, for example.) Having founders of the religion swear to the truth is pretty unconvincing.
The heart of the scientific method is trying to open your own assumptions to criticism.
However, Christian believers can't do that, because they think their eternal life depends on solid belief in their doctrine. Without abandoning their faith, they are not able to genuinely consider whether it is true or not.
His followers overturned the most powerful government in the world without use of force
"Without use of force" is an exaggeration.
By a parallel argument, do you think that (largely nonviolent) Buddhist evangelism of Asia proves that their beliefs are universally true? If not, what is the difference?
(btw, this article will probably be archived soon, but if you're finding the debate as interesting as i am please mail me.)
Nothing can be proven beyond any doubt, but evolution is pretty well established, as much as say the atomic theory of chemistry, or gravity. You may not want to accept it, but practically every unbiased (non-fundamentalist) person who examines the evidence will accept it.
I can't see any prima facie reason why it can't be tested. The origin of life is just an event that occurred at some time in the past. As with any past even we can try to understand it by trying to find and interpret evidence of the events, or by trying to reproduce a similar situation, or indeed just by eliminating impossibilities, amongst other methods.
A similar argument could have been made a few years ago about thunder being caused by electrons vs angry gods, or about the geocentric universe. What makes your case any different?
It's not reasonable to compare the likelihood of "an intelligent designer" vs a hypothesis of evolution from (say) ferrite clays, because they're not at comparable levels of detail and specificity.
To take your analogy of a crime scene: the police may have some trouble working out exactly which human committed the crime, and how. But that doesn't mean that we need to start assuming that the crime was committed by mythological beings. The general hypothesis ("it was done by a human") is well-established, even if the details ("it was done by the butler") are not yet.
At a very broad level of detail, we can weigh up "there is an intelligent designer", vs "there is no intelligent design but merely process". Some arguments can be made each way, but on the whole it seems unnecessary to introduce mysterious invisible actors without overwhelming evidence.
If you want to get specific, then mainstream science can provide a number of possible explanations for how various processes occur: life arrived on asteroids, evolved in clay, evolved in tide pools, etc. If you want to fairly challenge them, you ought to provide a countertheory at a comparable level of detail. Don't just say "God did it", but rather answer: By what mechanism did god move atoms around to create life? By what means did he associate souls and animals? Where did god come from? Why is your idea of god any more convincing than any of the thousands of others?
To be taken seriously, you have to answer all those questions in a way that is scientifically falsifiable. Creationists tend to intentionally leave the details fairly vague and hard to refute, saying that god did it through means that we can't possibly understand. That may or may not be true, but since it's not open to criticism or evaluation there is not much point in considering it.
Well, I laughed. :-)
So I suppose either they misunderstood/misremembered Knuth, or there is some positive aspect to it that Knuth can see and I cannot.
Or perhaps they were talking about MIX (the ~1970s processor) not MMIX (the modern RISC one)! Yes, now that I think about it that sounds like a more likely explanation -- it was probably just a typo. Self-modifying code for procedure returns might be much more representative of the old mainframe machines that Knuth was modelling in his original books. Given current programming languages and machine architectures (split I/D caches) it looks a bit perverse.
Of course, some versions of FORTRAN couldn't do recursive calls to a function, which would seem to be the problem with directly modifying the function's code to insert the return address.
... why not just store the address then?
I guess you could put a self-modifying trampoline on the stack containing the return address, but
No, that shouldn't help them.
Broadly, these filters look for "interesting" words: either ones that are often found in spam but not in nonspam, or vice versa.
The word "the" often occurs in both types; the word "slartibartfast" rarely in either. Therefore both of them are neutral.
The word "stop-on-solib-event" has never occurred in my spam yet, but if they started adding it (through a web robot) then eventually the filter would come to think that it was neutral, rather than a nonspam indicator as at present.
Interestingly, misspellings can be a really good indicator of spamminess. Many spams are sent repeatedly, either through multiple addresses, or because they're chain letters, or just because the stupid spammers send them repeatedly. If there's a characteristic misspelling that hasn't occurred anywhere else, it becomes a good way to identify the spam.
Another nice thing about bogofilter (and possibly others) is that it considers origin IPs and domains along with body words. So things like seed.net.tw are likely to be dumped unless they have some other strongly nonspammy words. For me, AOL and Yahoo have slightly spammy smells -- more because they're often forged than because much spam originates there, I think.
Eventually these filters may be defeated, but I think they will work well for a while.
I have a theory that the reason many people want to filter spam is not just because of bandwidth or time, but rather because the moronic writing and presentation is an insult to a thinking reader. It is quite literally junk mail. If by trying to get past these filters spammers have to act more like reasonable humans it's probably a good thing.
This is less of a blunt instrument than DNS blacklists and therefore probably a good thing.
Here's the point: if I develop confidence in bogofilter and build a good database, then I can use it as an ingress filter on some large (>8000 members) mailing lists that I run. Every spam that gets through there costs us bandwidth, and wastes the time of every subscriber. In addition, spam messages which do get through to the list typically generate additional traffic, either by people complaining about spam, or in bounces from recipients whose servers reject it.
So the return from a good spamstopper is potentially enormous.
Any messages which are false positives can be handled by some other means; e.g. forwarding to a postmaster for review.
I vaguely remembered that it was methane fully substituted with halogens. However I will admit that I googled to check. :-)
"The sooner you realize google is smarter than you, the smarter you are."
You know, dihydrogen oxide is pretty dangerous in high concentrations too. The length of a chemical name is a poor indication of its toxicitity.
More seriously, although Halon contains halogens (chlorine, fluorine, and bromine), so do most of the plastics used in a server room. You don't want to be around when they're burning.
Incidentally, this is why there are separate "plenum rated" ethernet cables: they're made from chemicals that won't break down into anything *too* nasty when they burn, and therefore they can legally be used in air vents and similar spaces.
Many people seem to believe this myth that halon removes oxygen from the air, and that it is deadly to humans. It's simply not true.
A fire requires three things: fuel, oxygen, and heat. A fire can be stopped in four ways: by removing fuel (e.g. shutting off a gas tap), by removing oxygen (e.g. smothering in CO2 or a fire blanket), by cooling (e.g. with a water hose), or by interrupting the chemical reaction. Halon is an example of this last strategy.
If you wanted to exclude oxygen from the machine room, the most economical and effective approach would probably be to dump CO2 into it. (This is why most office fire extinguishers use CO2 -- it works nicely on small fires and is safe around electrical equipment.) The problem is that in a confined space this would also asphyxiate any humans. (As another example of a confined space containing humans, halon is used on airplanes.)
Halon gas is used in concentrations of about 5%, at which level it is not harmful to humans for short periods. It's roughly as poisonous as other short organic molecules. It could be compared to the stuff you smell when using spraycans or working on a motor vehicle: you shouldn't do it all day or intentionally inhale it, but brief infrequent exposure is not particularly harmful. (Presumably your machine room doesn't catch fire every day...)
This is how it works, as I understand it: halon molecules compete with other chemicals in the room to preferentially absorb energy and free radicals from ignition points. This interrupts the chain reaction, so that energy liberated by burning goes into decomposing halon, rather than into setting more stuff alight. For this reason it can work at low concentrations.
The products produced when halon burns are bad for you, but as another poster pointed out they're no worse than other chemicals produced in a machine room fire.
Production of new halon systems has been tightly restricted because of their potential to damage the ozone layer. That says nothing about whether they're harmful for humans. (Indeed, the high UV flux that causes ozone breakdown in combination with CFCs would be pretty bad for you...)
Some links:
Chemistry of Halon
About Halon
Simple messages:
- Read and understand the MSDS (material safety data sheet) for chemicals in your workplace.
- If an uncontrolable fire breaks out, everybody should leave the room in an orderly fashion, close the door behind you, and call the fire service.
Regardless of the chance of conviction, perhaps Alan has better things to do than put himself through the stress, risk and expense of a criminal trial? The Sklyarov trial hardly inspires confidence in the US "justice" system.
Alan needs to turn over kernel development to someone with more balls than he has.
I don't see anyone stopping you from creating a fork... If you think your enormous testicles make you better qualified, please go ahead.
Yes, something like this is more general. distcc is only useful for compiling C and C++.
/home NFS-mounted, and a single software image -- then it's great.
However the drawback is that it requires you to use a shared filesystem and to have the same compilers and headers installed on all machines. If you already have that situation --
However many systems are not so tightly controlled. distcc is much less intrusive to install -- you can distribute jobs to a machine administered by somebody else without requiring them to start using NFS or stop upgrading libraries.
Yeah, linking is now the dominant factor in the Kernel Compile Wars. A 64-way POWER4 machine running 2.5.x can build it in about 4 seconds last time I checked, and most of that is linking.
A paralellizable linker might make a nice Masters project for someone.
The fact that ClearCase seems to slow down file operations by a factor of about 2 to 10 makes this argument not exactly compelling. I've seen it take 40 seconds to list a short directory, and on very well-hung machine.
Perhaps there are good things about ClearCase, but speed certainly isn't one of them.
That said, the function you described is basically what ccache does -- only without requiring stupid kernel tricks.