And what good exactly is WiFi going to do for countries where electricity is scarce and intermittent and computers are even more scarce?
If you want to give them a fishing rod, tell the US and EU to repeal their outrageous farm subsidies so the African farmers can actually get reasonable prices for their products.
Linus uses a very pragmatic philosophy - "the best tool for the job." Debian, by and large does not, and instead uses the Free Software Foundation's "the best tool for the job, as long as it's Free." Hence the Debian Free Software Guidelines, the name GNU/Linux, and the strong insistence on keeping the "non-free" portions of Debian optional and un-tied to the main part of the OS, which is wholly Free.
Furthermore, I'd argue that Debian is a good example of RMS's argument that Linux is just a kernel, and not the operating system. Debian is working on several alternate systems in which the major portion of the OS is kept the same, and the kernel is replaced. Thus Debian can be viewed as essentially a distribution of the GNU OS, with various versions including different kernels (hence Debian GNU/Linux, Debian GNU/HURD, and the in-discussion Debian GNU/BSD).
One of the more prominent and widely-used distributions is Debian GNU/Linux, which clearly ascribes to the GNU philosophy in designing its system, not the more pragmatic Linus philosophy.
You post on slashdot, so I am skeptical of your story involving sex and a girlfriend.
Unions are even worse
on
Working Hard?
·
· Score: 2, Insightful
If you're qualified, you don't want to join a union. Want to teach high school and have skills that are in demand for high school teachers (science and math)? Too bad, you'll get paid the same as the dime-a-dozen social studies teachers: union says so. Want to negotiate your own pay? Too bad: union membership (and paying union dues) is mandatory for employment. Etc.
Unions have a bad habit of discouraging merit. It doesn't matter if you have useful skills, or if you're a better employee than the others; all that matters is your seniority. This is why there are absolutely horrid teachers who are getting paid more than everyone else, because they've been there 40 years (oh, and you can't fire them either, unless they do something completely blatant like molest children: union says so).
also in contrast to europe
on
Working Hard?
·
· Score: 1
Stuff actually gets done in the US. I know plenty of people with American firms working with EU contractors, and the Europeans are always on a goddamn vacation. They take off something like 6-8 weeks/year on average (at least in the white-collar jobs), and always at the most inopportune times. And fairly often the entire office closes for "major holidays" which seem to happen dozens of times per year (by contrast, most largish American firms will rotate holidays so everyone isn't gone at once, except perhaps on Christmas Day).
It may be possible to replace the C library with the one from FreeBSD, but it would likely be non-trivial, and I'm not aware of it having been done -- glibc has been designed to work with the Linux kernel, whereas the *BSD C libraries haven't been.
The biggest culprit for FreeBSD's lack of popularity compared to Linux is probably not the license, but the unfortunate mess of lawsuits all the *BSDs were involved in in the early 1990s. After the resolution of these lawsuits, Linux had already started. FreeBSD was probably still ahead of Linux at this time, but the developers had already gained some momentum behind Linux and didn't want to switch. The BSD vs. GPL issue isn't really relevant there, because if they had wished a group of developers could've taken the FreeBSD kernel, modified it, and relicensed the modified version as GPL, and continued development from there (since the BSD license does not prohibit adding additional restrictions to derived versions, which is why you can use BSD-licensed code in GPL-licensed projects).
Due to licensing and ideological issues the free *BSDs have been trying to do this for years, but have not succeeded in replacing GNU components of their OS.
If you're willing to use non-Free software, you may be able to come a little closer, by replacing gcc with icc as in your example. Clearly this is unacceptable to those building Free operating systems, however.
GNU is in fact an OS without a kernel, not just some text editors. Perhaps its most important core-OS component is the GNU C library (glibc), which performs a similar function to the core system DLLs in Windows. There's also the GNU Compiler Collection (gcc) which is absolutely essential for actually using software, as there are no other Free compilers of which I'm aware (even Free/Open/NetBSD use this GPL'd compiler for that reason).
Frankly, the OS is more essentially GNU than Linux. As pointed out, it's possible to replace the Linux kernel and still have a working OS, albeit one that's quite a bit less advanced than currently. However, it's completely impossible to replace the GNU components in any timespan less than decades.
Usually the way these things work is by comparing against a giant database of "papers available on the web" and using various algorithms for judging similarity. This catches at least 95% of people who downloaded a paper from the web and made minor changes to it. In CS, where submissions are usually electronic, profs will often augment this database with all previous submissions for the course, sometimes dating back a decade or so. This catches at least 95% of people who copied an assignment from a friend who took the class a year or two ago (or even someone who graduated 5 years ago) and made minor changes (like search-replace on variable names, or some minor structural reworking) to it. Obviously this is harder to do in the humanities, unless you require electronic submissions in some format parsable as text or you OCR everything turned in.
As for unattributed quotes, you're certainly correct there. It's a completely intractable problem: the only way to know for sure that a particular sentence (or paragraph) was not plagiarized from somewhere is to check it against every single paragraph ever written in the history of the written word. Checking against some common sources might work decently though, especially if limited to a specific field (i.e. you can probably catch a significant percentage of plagiarized paragraphs in an anthropology paper by using a database of the 1000 most-cited anthropology books/papers).
But in any case, these things are mainly targetted at outright cheating: copying entire essays or large portions of essays from someone else.
Also had to say it...
on
Jaguar is Over
·
· Score: 4, Funny
I also don't know of any other business where the employees get to live like rock stars. It's not like most artists actually work 40-hour weeks like everyone else does. And the biggest problem, from the label's point of view, is that nobody wants to buy CDs from most of them. Thus, the label makes $0 on most of them, so has to take most of the profits from the few that do sell.
Sure, artists can skip labels. Some might make it big on their own: as a bigger example of success, Bad Religion started out putting out their own records under the then-non-existent "Epitaph Records" imprint. Most will not make it at all: they'll be the literally thousands of no-name bands littering mp3.com.
So while I'll agree record labels for the most part suck, the problem is that artists also for the most part suck.
yes, hundredths, and it's a kernel limitation
on
Sun's Last Stand
·
· Score: 1
The Linux kernel uses a 32-bit integer for counting ticks, and ticks go by at the rate set by the kernel constant HZ, which is currently 100 on x86 systems. 2.5.x has increased this to a 64-bit integer and increased HZ to 1000 (to offer better timing accuracy: right now Linux timers have a resolution of only 10ms, which is far inferior to Win2k/XP and unsuitable for audio-type work; though as an aside even 1ms timers are far worse than XP's 100us timers).
Just in case anyone isn't clear on why short-selling is bad if you don't have a lot of experience: there's no limit on how much money you can lose. With normal stock investment, if you invest, say, $1000 in stock, the worst you can do is lose $1000 (if the stock become worthless). You basically choose beforehand the upper bound on your potential losses.
With short-selling, on the other hand, there is no potential upper bound. If you short-sell SCO, it could go up to $100/share, it could go up to $1000/share, it could go up to $2.3m/share. While some of these prices are more likely than others, the point is that you can lose many times your initial investment.
This also makes it harder to make a profit. For you to make money, the stock can go down only by $x, where $x is its current price, but for you to lose money, the stock can go up by any amount. So the balance is tilted away from you.
It's even harder to maintain your code, 'cause now none of the line numbers are actually what the source file says they are. Nothing like having to mentally trace the execution of a program in order to figure out where your labels are.
It sounds like you're basically positioning your hands like on a split keyboard, only you're using a regular keyboard to do it. ASDF on a split keyboard is about at the same angle as QSDC on a standard keyboard.
There has been significant harrassment of Egypt's coptic population, especially by Islamists: this 6% used to be much higher, but many have left. I also noticed you left out the other of the "three divine religions", Judaism. There used to be a relatively large Jewish minority in Egypt, but they have since been entirely driven out.
Slashdot Mirror
Do those rates include calling anywhere within Europe, or only within the very small area of your own country?
And what good exactly is WiFi going to do for countries where electricity is scarce and intermittent and computers are even more scarce?
If you want to give them a fishing rod, tell the US and EU to repeal their outrageous farm subsidies so the African farmers can actually get reasonable prices for their products.
Writting up comment slashdot box into internet commentwise albeit did, however. To read is difficulty I fear.
Linus uses a very pragmatic philosophy - "the best tool for the job." Debian, by and large does not, and instead uses the Free Software Foundation's "the best tool for the job, as long as it's Free." Hence the Debian Free Software Guidelines, the name GNU/Linux, and the strong insistence on keeping the "non-free" portions of Debian optional and un-tied to the main part of the OS, which is wholly Free.
Furthermore, I'd argue that Debian is a good example of RMS's argument that Linux is just a kernel, and not the operating system. Debian is working on several alternate systems in which the major portion of the OS is kept the same, and the kernel is replaced. Thus Debian can be viewed as essentially a distribution of the GNU OS, with various versions including different kernels (hence Debian GNU/Linux, Debian GNU/HURD, and the in-discussion Debian GNU/BSD).
One of the more prominent and widely-used distributions is Debian GNU/Linux, which clearly ascribes to the GNU philosophy in designing its system, not the more pragmatic Linus philosophy.
You post on slashdot, so I am skeptical of your story involving sex and a girlfriend.
If you're qualified, you don't want to join a union. Want to teach high school and have skills that are in demand for high school teachers (science and math)? Too bad, you'll get paid the same as the dime-a-dozen social studies teachers: union says so. Want to negotiate your own pay? Too bad: union membership (and paying union dues) is mandatory for employment. Etc.
Unions have a bad habit of discouraging merit. It doesn't matter if you have useful skills, or if you're a better employee than the others; all that matters is your seniority. This is why there are absolutely horrid teachers who are getting paid more than everyone else, because they've been there 40 years (oh, and you can't fire them either, unless they do something completely blatant like molest children: union says so).
Stuff actually gets done in the US. I know plenty of people with American firms working with EU contractors, and the Europeans are always on a goddamn vacation. They take off something like 6-8 weeks/year on average (at least in the white-collar jobs), and always at the most inopportune times. And fairly often the entire office closes for "major holidays" which seem to happen dozens of times per year (by contrast, most largish American firms will rotate holidays so everyone isn't gone at once, except perhaps on Christmas Day).
"They should use Postgre because MySQL isn't a real database; I really need rollbacks and views in my app."
You just ruined 25 jokes in one post.
Multicast exists in IPv4 as well, but no ISPs support it since no one's worked out a pricing model for it yet.
It may be possible to replace the C library with the one from FreeBSD, but it would likely be non-trivial, and I'm not aware of it having been done -- glibc has been designed to work with the Linux kernel, whereas the *BSD C libraries haven't been.
The biggest culprit for FreeBSD's lack of popularity compared to Linux is probably not the license, but the unfortunate mess of lawsuits all the *BSDs were involved in in the early 1990s. After the resolution of these lawsuits, Linux had already started. FreeBSD was probably still ahead of Linux at this time, but the developers had already gained some momentum behind Linux and didn't want to switch. The BSD vs. GPL issue isn't really relevant there, because if they had wished a group of developers could've taken the FreeBSD kernel, modified it, and relicensed the modified version as GPL, and continued development from there (since the BSD license does not prohibit adding additional restrictions to derived versions, which is why you can use BSD-licensed code in GPL-licensed projects).
Due to licensing and ideological issues the free *BSDs have been trying to do this for years, but have not succeeded in replacing GNU components of their OS.
If you're willing to use non-Free software, you may be able to come a little closer, by replacing gcc with icc as in your example. Clearly this is unacceptable to those building Free operating systems, however.
GNU is in fact an OS without a kernel, not just some text editors. Perhaps its most important core-OS component is the GNU C library (glibc), which performs a similar function to the core system DLLs in Windows. There's also the GNU Compiler Collection (gcc) which is absolutely essential for actually using software, as there are no other Free compilers of which I'm aware (even Free/Open/NetBSD use this GPL'd compiler for that reason).
Frankly, the OS is more essentially GNU than Linux. As pointed out, it's possible to replace the Linux kernel and still have a working OS, albeit one that's quite a bit less advanced than currently. However, it's completely impossible to replace the GNU components in any timespan less than decades.
Usually the way these things work is by comparing against a giant database of "papers available on the web" and using various algorithms for judging similarity. This catches at least 95% of people who downloaded a paper from the web and made minor changes to it. In CS, where submissions are usually electronic, profs will often augment this database with all previous submissions for the course, sometimes dating back a decade or so. This catches at least 95% of people who copied an assignment from a friend who took the class a year or two ago (or even someone who graduated 5 years ago) and made minor changes (like search-replace on variable names, or some minor structural reworking) to it. Obviously this is harder to do in the humanities, unless you require electronic submissions in some format parsable as text or you OCR everything turned in.
As for unattributed quotes, you're certainly correct there. It's a completely intractable problem: the only way to know for sure that a particular sentence (or paragraph) was not plagiarized from somewhere is to check it against every single paragraph ever written in the history of the written word. Checking against some common sources might work decently though, especially if limited to a specific field (i.e. you can probably catch a significant percentage of plagiarized paragraphs in an anthropology paper by using a database of the 1000 most-cited anthropology books/papers).
But in any case, these things are mainly targetted at outright cheating: copying entire essays or large portions of essays from someone else.
Of course it's dead; it's based on *BSD.
I also don't know of any other business where the employees get to live like rock stars. It's not like most artists actually work 40-hour weeks like everyone else does. And the biggest problem, from the label's point of view, is that nobody wants to buy CDs from most of them. Thus, the label makes $0 on most of them, so has to take most of the profits from the few that do sell.
Sure, artists can skip labels. Some might make it big on their own: as a bigger example of success, Bad Religion started out putting out their own records under the then-non-existent "Epitaph Records" imprint. Most will not make it at all: they'll be the literally thousands of no-name bands littering mp3.com.
So while I'll agree record labels for the most part suck, the problem is that artists also for the most part suck.
The Linux kernel uses a 32-bit integer for counting ticks, and ticks go by at the rate set by the kernel constant HZ, which is currently 100 on x86 systems. 2.5.x has increased this to a 64-bit integer and increased HZ to 1000 (to offer better timing accuracy: right now Linux timers have a resolution of only 10ms, which is far inferior to Win2k/XP and unsuitable for audio-type work; though as an aside even 1ms timers are far worse than XP's 100us timers).
Just in case anyone isn't clear on why short-selling is bad if you don't have a lot of experience: there's no limit on how much money you can lose. With normal stock investment, if you invest, say, $1000 in stock, the worst you can do is lose $1000 (if the stock become worthless). You basically choose beforehand the upper bound on your potential losses.
With short-selling, on the other hand, there is no potential upper bound. If you short-sell SCO, it could go up to $100/share, it could go up to $1000/share, it could go up to $2.3m/share. While some of these prices are more likely than others, the point is that you can lose many times your initial investment.
This also makes it harder to make a profit. For you to make money, the stock can go down only by $x, where $x is its current price, but for you to lose money, the stock can go up by any amount. So the balance is tilted away from you.
It's even harder to maintain your code, 'cause now none of the line numbers are actually what the source file says they are. Nothing like having to mentally trace the execution of a program in order to figure out where your labels are.
It sounds like you're basically positioning your hands like on a split keyboard, only you're using a regular keyboard to do it. ASDF on a split keyboard is about at the same angle as QSDC on a standard keyboard.
There has been significant harrassment of Egypt's coptic population, especially by Islamists: this 6% used to be much higher, but many have left. I also noticed you left out the other of the "three divine religions", Judaism. There used to be a relatively large Jewish minority in Egypt, but they have since been entirely driven out.
I would've paid $10 for a reformatted version, but now I've stolen it instead.