No, that's the usual kind of picture that you see in Wall Street Journal articles. Go down to your local news-stand or library and look at a couple of Journals and you'll see.
That's right. Marathon runners are vulnerable to hyponatremia. Massive sweat + intake of unsalted water leads to sodium ion imbalance inside the body. It's one of those nasty conditions where the brain gets disoriented so the victim doesn't realize that they are headed for death.
I know the Slashdot stereotype is that nobody *here* has to worry about such things, but actually, I bet there are people in the Slashdot community who run this far and this hard.
Canopy Group (parent of SCO) and Center 7, another Canopy subsidiary, had a joint marketing arrangement with CA. Canopy claims that CA welshed. Canopy and Level 7 sued CA. The suit was settled with a $40 million payment.
I seem to recall, but I can't find a link, that other terms of the suit were that CA buy some Linux licenses. That would fit in with Canopy's plans.
Good point, I hadn't thought about how many election days there are.
As far as I know, there aren't any significant groups of people who have a problem with voting on Mondays, so moving election days from Tuesday to Monday would be okay if it were part of a larger plan.
Some people, such as Orthodox Jews, restrict their activities on Saturday. You might reply "tough for them", but any change that makes voting harder for a significant class of people is going to be opposed by elected office-holders from any party that draws support from Saturday-observing people. That's why this proposal won't go anywhere in the U.S.
Here's a different proposal: make Election Day a national holiday. A lot of people would also take the Monday off as well. I think that democratic elections are important enough to be a national holiday, don't you?
The UAW (United Auto Workers union) negotiated a contract where Election Day is a paid holiday for their members. Good for them.
Computers are Turing-Complete machines. That is to say, any computation that can be done by any sort of device can be done by a modern computer.
Security is about denying completeness. For example, the Java security model, and the Unix "chroot jail", and a network firewall are all based on forbidding certain things that are possible in a Turing-complete environment. Conversely, Microsoft is determined to ship Turing-complete mailers, which are then prone to virus infection.
The problem with counting votes is not writing the code "for $precinct in @precinct_list; $vote{candidate} += $precinct_vote{$candidate}". The problem is *proving* that nobody else has added "$vote{MY_FAVORITE_CANDIDATE} += 1000" to the same program.
And there are plenty of people out there, on all sides of the political spectrum, who would cheat like that.
The same John Rechy who wrote "City of Night" and "The Sexual Outlaw" ?
I guess I shouldn't be surprised that he was hustling reviews on Amazon the same way that his characters hustled and tricked their ways through his books.
Rechy's books disturbed me, which is a good quality in a book. How can I describe them for a Slashdot crowd? Start with William Gibson or Bruce Sterling; subtrace all the "cyber" part of "cyberpunk"; and replace it with gay sex. LOTS of gay sex.
Customers like it when prices go down and quality goes up. That's the way a market is supposed to work.
Software sellers won't make as much money in a marketplace with open source software, but software buyers will get better products at lower prices. Even the software buyers who don't choose open source software will benefit from the increased competition among sellers.
I'm probably going to die, and so are you. In fact, no human being has ever reached 2^16 days of life.
Face it: your life is a finite resource. There are things you can do to conserve and extend that resource. But at the end of each day, you've traded that day for whatever love or money or experience or creative work or good deads or hedonism that you chose.
Plenty of people make commitments longer than a day. If you go to school, you trade years of your life in exchange for knowledge and socialization. If you play music, you spend a lot of time practicing. If you want to be a doctor or a politician or a writer or just about anything worthwhile, that's years of experience that you'll need to accumulate.
So now we're talking about a life commitment. It's just a bigger scale. Whether it's worthwhile or not is an empirical decision which properly belongs to the individual who decides how to commit their life.
YYYY-MM-DD and YYYYMMDD are both standards-compliant.
Seriously, if you've never heard of this standard, read up. Whenever I need to stick a date or a time on something in text form, I just do it the ISO 8601 way.
paypal, ebay, circuit city, bank of america, and microsoft all have reason to publish SPF records.
Criticizing Linux on Slashdot
on
What You Can't Say
·
· Score: 2, Interesting
Slashdot does have biases. But when one of those hot-button topics comes up, I'll see a bunch of +5's on the minority side, as well.
It's a continuum. On the one hand, try going into a PC environment and talking about race&intelligence and see how fast people will literally shun you. Or a conservative group, and talk about gay marriage.
On the other hand, next time a "Linux rulez/sucks" thread pops up, try posting some thoughtful pro-Windows comments, and see if people respond to the actual points you make, or just knee-jerk. I really think Slashdot is pretty good on the rational debate.
I don't remember Asimov mentioning the race of the major characters in I, Robot. But he does mention the races of some minor characters in "The Evitable Conflict", such as Lincoln Ngoma.
Ngoma laughed. He was a big, dark man, strong faced and handsome.
So, do you admit that Linus copied the entire substance of include/errno.h from somewhere else?
I dug out Unix 5th Edition because it's the closest to the original that I could find. Linus says that he copied his errno.h from Minix. Tanenbaum probably copied his from somewhere else. The chain of copying goes back to the original AT&T Unix code.
Compare with the paranoid attitude of gcc: gcc doesn't include copies of other people's system header files from, say, solaris or hp-ux; gcc includes scripts that manipulate the files on the end user's system, if the end user has legal copies of these files.
RMS has a stick up his a** about a lot of things, but better RMS's stick than Boies' dick.
I haven't read the Posix standard...
Here is the 2003 version of the standard for errno.h.
The names are standardized in 2003 (and were probably standardized before 1991). But the values are not standardized. If you or I were to write an errno.h from that specifcation, the odds are incredible that we would happen to choose the same 31 values in the same order as the original Unix implementation. Personally, I would follow the order in the specification, which is alphabetical.
Again, from a legal point of view, it's a molehill. The actual value of those identical error numbers numbers is not $3,000,000,000. More like $300 or $3000. Ditto with the ioctl numbers. As you point out, the values of certain signal numbers (like SIGKILL) are well known. And all of these files are completely unrelated to SCO's claims that IBM copied SMP, NUMA, JFS, RCU code into Linux.
But from a PR point of view, it's damaging to us. This activity looks uncomfortably like SCO's allegations about how Linux is developed. Here's Unix code; here's Linux code; the Linux code is a verbatim copy of the Unix code.
A nasty twist: because this is part of the ABI, it would be difficult and painful to replace include/asm-*/errno.h with a clean room implementation from the Single Unix Specification.
A note here: I'm not pro-SCO, and I'm not trolling. I'm looking to find the strongest SCO points because we're at war with SCO, so it's important to understand the strong points of the enemy's position.
Okay, mods, get out the mod-bombs like you did with the parent. But first, read my other comments this year -- I've put in a lot of time digging into SCO financials and reporting on SCO conference calls.
Start here: google "sco ancient unix". Follow the link. Download the tarball for 5th Edition UNIX, v5root.tar.gz. Unpack it. Look at the list of error numbers at the end of usr/sys/user.h. Note that the date on usr/sys/user.h is 1974-11-26.
Then download linux-0.0.1.tar.bz2 from ftp.kernel.org. Unpack it. Look at include/errno.h. Notice the literal copying of values.
This file is not an original work by Linus Torvalds. He says as much in the comment which the parent poster quoted.
Of course, there are questions: did Linus copy this file legally from a public-domain or freely licensed source, such as a standards body? Even if the copying is illegal, has SCO been damaged? Has SCO sent proper specific notice to people and corporations that distribute errno.h? Now, go ahead and mod-bomb me the way you did the parent. I've spent a ton of time reseasching SCO and posting the results of my research here and on Groklaw. But check the references I quoted before you moderate. Are you moderating because my argument is faulty, or are you moderating because you don't want to hear the strong points of the enemy's case?
Okay, googled on "SCO Ancient Unix" and found a nifty mirror here.
I'm looking at usr/sys/user.h from Unix V6 in one window: #define EPERM 1, #define ENOENT 2, #define ESRCH 3, #define EINTR 4, and so on. And I'm looking at include/asm-i386/errno.h from linux 2.4.20 in another window: #define EPERM 1, #define ENOENT 2, #define ESRCH 3, and so on.
Several posters have pointed to the Open Group spec. As you say, the names are part of the standard. It's also necessary that the names be identical for compatibility. Linus had no choice but to write "#define ESRCH...". "#define ESEARCH" or "#define EPROCNOTFOUND" would not work.
However, the IEEE Std 1003.1 spec does not list the numeric values. In fact, the 2003 edition that I'm looking at lists the names alphabetically: E2BIG, EACCESS, EADDRINUSE, EADDRNOTAVAIL...
There are 32 error code numbers in the V6 error number list. 31 of them are identical in Linux 2.4.20. In my opinion, SCO has offered credible evidence that the files are substantially similar. Now the burden is on Linus to show that he copied ENXIO=6, ENOTDIR=20, ENFILE=23, and so on, from a legitimate source.
(I agree with you about the extent of damages suffered by SCO -- nearly none. And IBM certainly didn't contribute any of this code to Linux.)
I wanna see the Wachowski brothers make "The Menace From Earth".
No, I don't wanna see Jeff in a trenchcoat and Holly in black PVC. I wanna see Ariel falling in bullet time with Holly chasing her. And I want the soundtrack to be QUIET while they are doing it.
The original poster claimed: "you have no inherent right to restrict what other people copy." I claimed: yes I do, I have a right to restrict what my doctor copies. Now you come in and agree with me that yes, I do have a right to restrict what my doctor copies.
Copyright has nothing to do with privacy...
Yes, it does. Both copyright and medical privacy are rights to restrict what other people copy. If you universally deny a right to restrict what other people copy, then you deny both copyrights and medical privacy.
You are saying that you want to restrict what other people copy in some cases ("private information") but not others ("played on the radio and sold in shops"). That's a point of view, but it's a different point of view from the original poster, who believes there should be no restrictions on doctors sharing medical information.
Here's a question for you: under your point of view, did Linksys do anything wrong when they picked up the Linux kernel, ported it to their devices, and sold the devices with binaries and without sources? The Linux kernel is surely readily available.
Slashdot Mirror
He doesn't live on the Upper West Side; he lives in "Midtown Manhattan", which is just as pricy.
You're probably right about the rent control. He probably got a sweetheart deal on the apartment.
No, that's the usual kind of picture that you see in Wall Street Journal articles. Go down to your local news-stand or library and look at a couple of Journals and you'll see.
(I think I've just been trolled).
That's right. Marathon runners are vulnerable to hyponatremia. Massive sweat + intake of unsalted water leads to sodium ion imbalance inside the body. It's one of those nasty conditions where the brain gets disoriented so the victim doesn't realize that they are headed for death.
Hyponatremia a Concern for Marathon Runners
I know the Slashdot stereotype is that nobody *here* has to worry about such things, but actually, I bet there are people in the Slashdot community who run this far and this hard.
New WTC Design
Actually, there's a set of buildings in Brooklyn which have exactly this profile (only smaller) when seen from the harbor.
Google this: "canopy ca settlement"
Canopy Group (parent of SCO) and Center 7, another Canopy subsidiary, had a joint marketing arrangement with CA. Canopy claims that CA welshed. Canopy and Level 7 sued CA. The suit was settled with a $40 million payment.
I seem to recall, but I can't find a link, that other terms of the suit were that CA buy some Linux licenses. That would fit in with Canopy's plans.
Link to the settlement
Good point, I hadn't thought about how many election days there are.
As far as I know, there aren't any significant groups of people who have a problem with voting on Mondays, so moving election days from Tuesday to Monday would be okay if it were part of a larger plan.
Multi-day elections have their own problems, namely the physical security of the polling equipment overnight.
...)
(And don't get me started on absentee ballots
Some people, such as Orthodox Jews, restrict their activities on Saturday. You might reply "tough for them", but any change that makes voting harder for a significant class of people is going to be opposed by elected office-holders from any party that draws support from Saturday-observing people. That's why this proposal won't go anywhere in the U.S.
Here's a different proposal: make Election Day a national holiday. A lot of people would also take the Monday off as well. I think that democratic elections are important enough to be a national holiday, don't you?
The UAW (United Auto Workers union) negotiated a contract where Election Day is a paid holiday for their members. Good for them.
Computers are Turing-Complete machines. That is to say, any computation that can be done by any sort of device can be done by a modern computer.
Security is about denying completeness. For example, the Java security model, and the Unix "chroot jail", and a network firewall are all based on forbidding certain things that are possible in a Turing-complete environment. Conversely, Microsoft is determined to ship Turing-complete mailers, which are then prone to virus infection.
The problem with counting votes is not writing the code "for $precinct in @precinct_list; $vote{candidate} += $precinct_vote{$candidate}". The problem is *proving* that nobody else has added "$vote{MY_FAVORITE_CANDIDATE} += 1000" to the same program.
And there are plenty of people out there, on all sides of the political spectrum, who would cheat like that.
The same John Rechy who wrote "City of Night" and "The Sexual Outlaw" ?
I guess I shouldn't be surprised that he was hustling reviews on Amazon the same way that his characters hustled and tricked their ways through his books.
Rechy's books disturbed me, which is a good quality in a book. How can I describe them for a Slashdot crowd? Start with William Gibson or Bruce Sterling; subtrace all the "cyber" part of "cyberpunk"; and replace it with gay sex. LOTS of gay sex.
Customers like it when prices go down and quality goes up. That's the way a market is supposed to work.
Software sellers won't make as much money in a marketplace with open source software, but software buyers will get better products at lower prices. Even the software buyers who don't choose open source software will benefit from the increased competition among sellers.
I'm probably going to die, and so are you. In fact, no human being has ever reached 2^16 days of life.
Face it: your life is a finite resource. There are things you can do to conserve and extend that resource. But at the end of each day, you've traded that day for whatever love or money or experience or creative work or good deads or hedonism that you chose.
Plenty of people make commitments longer than a day. If you go to school, you trade years of your life in exchange for knowledge and socialization. If you play music, you spend a lot of time practicing. If you want to be a doctor or a politician or a writer or just about anything worthwhile, that's years of experience that you'll need to accumulate.
So now we're talking about a life commitment. It's just a bigger scale. Whether it's worthwhile or not is an empirical decision which properly belongs to the individual who decides how to commit their life.
I got your international standard right here.
YYYY-MM-DD and YYYYMMDD are both standards-compliant.
Seriously, if you've never heard of this standard, read up. Whenever I need to stick a date or a time on something in text form, I just do it the ISO 8601 way.
paypal, ebay, circuit city, bank of america, and microsoft all have reason to publish SPF records.
Slashdot does have biases. But when one of those hot-button topics comes up, I'll see a bunch of +5's on the minority side, as well.
It's a continuum. On the one hand, try going into a PC environment and talking about race&intelligence and see how fast people will literally shun you. Or a conservative group, and talk about gay marriage.
On the other hand, next time a "Linux rulez/sucks" thread pops up, try posting some thoughtful pro-Windows comments, and see if people respond to the actual points you make, or just knee-jerk. I really think Slashdot is pretty good on the rational debate.
I don't remember Asimov mentioning the race of the major characters in I, Robot. But he does mention the races of some minor characters in "The Evitable Conflict", such as Lincoln Ngoma.
Ngoma laughed. He was a big, dark man, strong faced and handsome.
Rossum's Universal Robots
No kidding. Karel Capek invented the word "Robot" to refer to a class of genetically engineered worker slaves.
So, do you admit that Linus copied the entire substance of include/errno.h from somewhere else?
...
I dug out Unix 5th Edition because it's the closest to the original that I could find. Linus says that he copied his errno.h from Minix. Tanenbaum probably copied his from somewhere else. The chain of copying goes back to the original AT&T Unix code.
Compare with the paranoid attitude of gcc: gcc doesn't include copies of other people's system header files from, say, solaris or hp-ux; gcc includes scripts that manipulate the files on the end user's system, if the end user has legal copies of these files.
See this bit from the GNU coding standard:
Referring to Proprietary Programs
RMS has a stick up his a** about a lot of things, but better RMS's stick than Boies' dick.
I haven't read the Posix standard
Here is the 2003 version of the standard for errno.h.
The names are standardized in 2003 (and were probably standardized before 1991). But the values are not standardized. If you or I were to write an errno.h from that specifcation, the odds are incredible that we would happen to choose the same 31 values in the same order as the original Unix implementation. Personally, I would follow the order in the specification, which is alphabetical.
Again, from a legal point of view, it's a molehill. The actual value of those identical error numbers numbers is not $3,000,000,000. More like $300 or $3000. Ditto with the ioctl numbers. As you point out, the values of certain signal numbers (like SIGKILL) are well known. And all of these files are completely unrelated to SCO's claims that IBM copied SMP, NUMA, JFS, RCU code into Linux.
But from a PR point of view, it's damaging to us.
This activity looks uncomfortably like SCO's allegations about how Linux is developed. Here's Unix code; here's Linux code; the Linux code is a verbatim copy of the Unix code.
A nasty twist: because this is part of the ABI, it would be difficult and painful to replace include/asm-*/errno.h with a clean room implementation from the Single Unix Specification.
A note here: I'm not pro-SCO, and I'm not trolling. I'm looking to find the strongest SCO points because we're at war with SCO, so it's important to understand the strong points of the enemy's position.
Okay, mods, get out the mod-bombs like you did with the parent. But first, read my other comments this year -- I've put in a lot of time digging into SCO financials and reporting on SCO conference calls.
Start here: google "sco ancient unix". Follow the link. Download the tarball for 5th Edition UNIX, v5root.tar.gz. Unpack it. Look at the list of error numbers at the end of usr/sys/user.h. Note that the date on usr/sys/user.h is 1974-11-26.
Then download linux-0.0.1.tar.bz2 from ftp.kernel.org. Unpack it. Look at include/errno.h. Notice the literal copying of values.
This file is not an original work by Linus Torvalds. He says as much in the comment which the parent poster quoted.
Of course, there are questions: did Linus copy this file legally from a public-domain or freely licensed source, such as a standards body? Even if the copying is illegal, has SCO been damaged? Has SCO sent proper specific notice to people and corporations that distribute errno.h?
Now, go ahead and mod-bomb me the way you did the parent. I've spent a ton of time reseasching SCO and posting the results of my research here and on Groklaw. But check the references I quoted before you moderate. Are you moderating because my argument is faulty, or are you moderating because you don't want to hear the strong points of the enemy's case?
Hey, I've got user.h right here, from version 5 unix:
bash-2.05b$ ls -l usr/sys/user.h
-rw-r--r-- 1 mec mec 1217 Nov 26 1974 usr/sys/user.h
There's no copyright notice in the file. This is the file that defines all the errno's.
Same story with usr/sys/param.h (signal numbers). I dunno about ioctl's. They weren't in version 5 unix.
Okay, googled on "SCO Ancient Unix" and found a nifty mirror here.
...". "#define ESEARCH" or "#define EPROCNOTFOUND" would not work.
...
I'm looking at usr/sys/user.h from Unix V6 in one window: #define EPERM 1, #define ENOENT 2, #define ESRCH 3, #define EINTR 4, and so on.
And I'm looking at include/asm-i386/errno.h from linux 2.4.20 in another window: #define EPERM 1, #define ENOENT 2, #define ESRCH 3, and so on.
Several posters have pointed to the Open Group spec. As you say, the names are part of the standard. It's also necessary that the names be identical for compatibility. Linus had no choice but to write "#define ESRCH
However, the IEEE Std 1003.1 spec does not list the numeric values. In fact, the 2003 edition that I'm looking at lists the names alphabetically: E2BIG, EACCESS, EADDRINUSE, EADDRNOTAVAIL
There are 32 error code numbers in the V6 error number list. 31 of them are identical in Linux 2.4.20. In my opinion, SCO has offered credible evidence that the files are substantially similar. Now the burden is on Linus to show that he copied ENXIO=6, ENOTDIR=20, ENFILE=23, and so on, from a legitimate source.
(I agree with you about the extent of damages suffered by SCO -- nearly none. And IBM certainly didn't contribute any of this code to Linux.)
I saw a DOS 6.22 screen this week. It was a failed bootup on a New York City MTA card-vending machine.
I wanna see the Wachowski brothers make "The Menace From Earth".
No, I don't wanna see Jeff in a trenchcoat and Holly in black PVC. I wanna see Ariel falling in bullet time with Holly chasing her. And I want the soundtrack to be QUIET while they are doing it.
Well, if you don't like the new law, send an e-mail to every member of parliament and tell them what you think!
...
Oh wait
You haven't pointed out any flaws at all.
...
The original poster claimed: "you have no inherent right to restrict what other people copy." I claimed: yes I do, I have a right to restrict what my doctor copies. Now you come in and agree with me that yes, I do have a right to restrict what my doctor copies.
Copyright has nothing to do with privacy
Yes, it does. Both copyright and medical privacy are rights to restrict what other people copy. If you universally deny a right to restrict what other people copy, then you deny both copyrights and medical privacy.
You are saying that you want to restrict what other people copy in some cases ("private information") but not others ("played on the radio and sold in shops"). That's a point of view, but it's a different point of view from the original poster, who believes there should be no restrictions on doctors sharing medical information.
Here's a question for you: under your point of view, did Linksys do anything wrong when they picked up the Linux kernel, ported it to their devices, and sold the devices with binaries and without sources? The Linux kernel is surely readily available.