I'd really be interested to hear what you think the pro's and cons are with the current linux methodology. I guess I'm having trouble understanding why the ease I experience with windows programs would somehow be unfit for Linux.
You've named the Windows pros very well, so I'll just bring in a couple of Windows cons and the Linux pros. Keep in mind, I agree with everything you've said:).
Windows Cons
Most programmers suck. Most programs suck. They install stuff where it doesn't belong. This isn't Microsoft's fault, and everyone knows Microsoft doesn't encourage it, but many programs install stuff to the wrong place, and don't uninstall properly. This situation is getting better... except for spyware, which is getting more and more popular.
Every program adds a whole menu to the Start Menu, and lots of programs add icons to the Desktop and the Quick Launch folder. You also get system tray additions and programs that auto-load on startup. Not to mention, many add context-menu options in Explorer. Install ten programs and suddenly you have a very confusing UI, unless you spend ages micro-managing.
Linux pros (provided you do things the way your distro maintainers would recommend)
Most new package installations add maybe 0 or 1 menu entry to your Applications menu, in the proper place.
Packages are available from central repositories: you should not have to scour the Internet to find them. Not to mention, they're free. Open up Synaptic and you can search for any package and install/remove it.
Open source means no spyware, adware, or any other such disgusting nonsense.
I prefer Linux for these reasons. In Windows all these "Linux pros" are obviously possible, but the point is, you have to do them manually. That gets to be a pain in the neck. To install a program on Windows, I'd have to Google for the file, download it (beware of counterfeit files and spyware), reorganize my Start menu, delete a Quick Launch and Desktop icon, run Regedit and delete entries from CurrentVersion/Run, set all the options on the program (which default to what the programmer wants you to use, not what you want to use), etc. In Linux, there's no hassle.
Compare two similar programs: install KaZaA on Windows, or giftui on Debian (note: gift-fasttrak isn't part of Debian's repositories; it's still easier to install, though).
I do agree with all your pros and cons; however, I will always prefer Linux for the reasons mentioned above.
Goddammit I hate this one. A TON OF LITTLE PROGRAMS ARE DISTRIBUTED AS EITHER SOURCE OR ONE OF THE FORMATS LISTED.
Then install to/usr/local or/opt, and don't mess up the operating system your distributors have worked so hard to perfect.
Better yet, package the files properly yourself. It's not rocket science. Or you can submit a project recommendation to your distro's wishlist.
Just because the program isn't available for your particular distribution, don't screw up your system by installing packages that weren't made for it. It's just asking for trouble.
Do not carry the Windows philosophy over to Linux. If you do not understand the fundamental concepts behind package managers, do not mess around with them. Most of them do their jobs extremely well unless you decide to try and do their jobs for them.
So the point is, we need something equivalent to "Add/Remove Programs" that just *works* on all linux distros.
Wrong. That's something a Windows user just beginning a migration to Linux would say.
Distributions are different. Full stop. If they were all the same, there would be no point in having several. But assume there is a point in having different distributions, because if there isn't then your sentence is wrong in an even more fundamental manner.
How are distributions different? They're different in where they place their files, how they split up the same source tarball into different packages, what package manager they use, etc. If you take the binary files from a Red Hat.rpm and try using them on Debian, you're on your own, even if your architectures are the same, because the configuration will most likely not be the same as what a Debian maintainer would use.
Long story short: do not install a.rpm on Debian, don't install a SuSE.rpm on Fedora, don't install a.deb on SuSE, etc. If you do, you're just asking for trouble.
There are already Linux equivalents of Windows' Add/Remove Programs (minus the godawful interface). Synaptic, YaST, etc.
In short, there is no problem here. Work on more important stuff (such as integration on the desktop).
During daylight savings you have to use 1, not 12.
But that's ignoring a bigger problem: the "halfway" point is really halfway when the sun's path is 45 degrees away from you. During summertime you'll have to bias yourself to look closer to the 12, and during wintertime farther from it.
In fact, the only place where the rule is universally true is the North Pole, since every direction is South. But there's another problem: the trick will only work for half the year.
I was excited to look at these benchmarks because I know that Linux does 64 bits and I really, really want to see what kind of difference there is before buying an Athlon64. But there is no 64-bit testing. What is the point of running on Linux if you don't take advantage of what Linux offers?
Quality Control does not necessarily mean "user interface". You can make a very high quality cell phone that still sucks. Nokia has an entire factory full of them.
I used to have a Nokia 3390 and it even though it wasn't perfect, it wasn't far off. One big button does what you expect to do all the time. So you just press the big button a few times, and it's all logical.
But that phone died, so I got a Siemens M55. Warning to anybody shopping for a phone: do not get a Siemens M55. Of all the useless features, this one has useless blinking lights on the sides of the phone -- and they constantly blink by default! The confirmation dialogs don't have a consistent placement for Yes/No and the games take ages to load. But that's not what turned me off this phone -- the fact of the matter is, I can't tell when people are calling me! The vibrator is weak and the ringer is quiet. If I'm walking down the street and somebody calls, I'll miss the call.
I'm returning it for a Sony Ericcson T300. I hope it's not just the same story all over again, but I suspect it will be.
All I want is a small, usable phone which also does SMS.
Who comes up with these numbers? Obviously somebody severely misinformed or dishonest. Who would benefit?
A totally incompetent sysadmin (trusted by his boss)
An anti-virus or security firm
The US government? ("Look, it's terrorism!")
A lousy techie looking to impress his girlfriend
Some drunk bum
The truth is, it's practically free to secure against any new virus. Once your system is reasonably secure you'll only incur maintenance costs, almost never incident costs.
To NVidia's credit, they seem to be somewhat serious about supporting Linux in a somewhat timely manner.
It feels to me like nVidia is a big corporate giant, bending down to let the OSS community eat out of its grubby paws whatever scraps it leaves them.
There's not "always" something to complain about. There's only one thing to complain about, and that's that nVidia does not co-operate with the Linux community. If it did, nVidia drivers would be in the kernel tree and it'd be twice as easy for all of us to get our cards working on a do-it-yourself kernel. Not to mention, we'd have fewer bugs and probably faster framerates.
Linux would benefit. nVidia would benefit.
Drivers are commodities, there's really no good reason to keep them closed. ESR has written plenty of convincing material on the matter, so I won't re-hash it.
[Student teacher trust is] being violated so regulary by students cheating that teachers wonder if recent degreed graduates really learned anything.
As a student of McGill, I can say quite confidently that no student in arts can graduate without learning a hell of a lot, cheating or not. As for science and engineering, teachers accept as a matter of course that students will collaborate, so assignments are worth very little (i.e., 10%-20% of the grade) and the rest is midterms and final exams (ignoring the issue of cheating on final exams).
I mean, really. A student can't plagiarize a 20-page paper without getting caught, right? But for a 5-page paper it might be possible. The student would either: a) submit something downloaded off the Internet, which is very easy to detect, or b) modify something off the Internet to make it hard to recognize -- thereby learning about the subject, because it would take a lot of work not to get caught.
So I think any student who plagiarizes without getting caught is doing a hell of a lot of work anyway, and is learning. Which is good, because you can't quite plagiarize on midterms and final exams, which account for most of the marks in most courses (even arts).
CVS, like most version control systems, has the concept of "branches"
This is just one example of the article's condescension. Duh. CVS, like most version control systems, also sucks enormously. The Linux kernel development process has the concept of "branches" too (and has had it long before Bitkeeper came around).
In fact, the entire article can basically be summed up in one sentence understandable to the entire target audience:
"BSD is an entire distribution maintained by the same people who hack at its base."
We all know that Netscape 4 is an awful, crashy, buggy, standards-breaking piece of crap that set the Internet back years.
Yes, it was an awful, crashy, buggy, standards-breaking piece of crap; but it certainly wasn't the awful, crashy, buggy, standards-breaking piece of crap which set the Internet back years -- and is still doing so.
The GPL states that you only have to provide the sources along with the distibution. The distibution is in this case embedded in their player, so the only thing they would have to answer to is a demand from a verified owner of the player.
Or a demand from anybody who downloaded the ROM from their website, such as the MPlayer developers.
Whoever picked the managers, and supervised the managers, is as much to blame as the damn foam chunk.
I'd say they're entirely to blame. It's happened before, too, with the Challenger disaster. Management was told that Challenger would explode, but they launched anyway.
With the Columbia disaster, critical safety measures were being overlooked. Concerned engineers asked three times for satellite photographs of Columbia, and never got them. Management knew about the foam problem for ages but never helped fix it.
The article basically is correct in stating that passionate dissagreements fork projects. The doubling up of energies on very similar projects (like Gnome and KDE) work against open source.
GNOME and KDE are not forks of the same project. But I'd agree that the general rule of thumb with forks still applies to them:
Forks are healthy.
Let's back it up with examples: egcs/gcc; XFree86/Xouvert/freedesktop.org; samba/samba-tng; apache 1.x/2.x; galeon/epiphany... the list goes on.
I think the examples show that forking is very healthy for open-source projects. In fact, I can't even think of a counterexample. Is there a single case in which a fork resulted in lower-quality code or stagnant development on both branches?
"I am terrified of the following scenario, which is extremely probable to happen soon," responded Jan Rychter in a Wednesday post. "2.4 is being moved into 'pure maintenance' mode and people are being encouraged to move to 2.6. While people slowly start using 2.6, Linus starts 2.7 and all kernel developers move on to the really cool and fashionable things. 2.6 bug reports receive little attention, as it's much cooler to work on new features than fix bugs."
What insight! I can't imagine how anybody could see that far into the future as Jan Rychter.
So many people use the "cooler to work on new features" argument without evidence. Since when has Linux favoured new features over fewer bugs? Linux 2.4 bugs receive plenty of attention, thank you very much -- so do Linux 2.2 bugs and 2.0 bugs.
Heh. My apartment has two 25-amp circuits -- appliances included.
My three computers are on the same circuit as the fridge, microwave, toaster and popcorn popper. If we run any three (or sometimes even two) of these appliances at the same time, the fuse blows.
Blown fuses are especially common during the summer, when I have to add on an air conditioner.
I couldn't help but notice, this guy put "WebCT" and "quality software" in the same sentence. Just because the software (online teaching suppliment) sells a lot doesn't make it good. I've seen a hell of a lot of bad free software, but never as bad as WebCT.
I'm having to reboot back to 2.4 for the Cisco vpn driver
Though MPPE (weak Microsoft encryption) is not available for 2.6 *yet*, you might want to look into pptpclient. The web page tells you exactly how to set up ppp and everything. It'll connect to pretty much any Windows-ish (i.e., Cisco) VPN, and it's free. MPPE support needs a kernel patch for 2.4 and is on its way for 2.6.
Advantages: GPL, better documentation, the primary developer is very helpful on the mailing list. Disadvantage: maybe harder to set up (not sure, never tried the Cisco one).
I'll bet 2.6 will have MPPE support built in before Cisco releases a 2.6 version of its VPN client.
Slashdot Mirror
I've been using subversion on Debian unstable for months. Just apt-get install subversion.
I'd really be interested to hear what you think the pro's and cons are with the current linux methodology. I guess I'm having trouble understanding why the ease I experience with windows programs would somehow be unfit for Linux.
You've named the Windows pros very well, so I'll just bring in a couple of Windows cons and the Linux pros. Keep in mind, I agree with everything you've said :).
Windows Cons
Linux pros (provided you do things the way your distro maintainers would recommend)
I prefer Linux for these reasons. In Windows all these "Linux pros" are obviously possible, but the point is, you have to do them manually. That gets to be a pain in the neck. To install a program on Windows, I'd have to Google for the file, download it (beware of counterfeit files and spyware), reorganize my Start menu, delete a Quick Launch and Desktop icon, run Regedit and delete entries from CurrentVersion/Run, set all the options on the program (which default to what the programmer wants you to use, not what you want to use), etc. In Linux, there's no hassle.
Compare two similar programs: install KaZaA on Windows, or giftui on Debian (note: gift-fasttrak isn't part of Debian's repositories; it's still easier to install, though).
I do agree with all your pros and cons; however, I will always prefer Linux for the reasons mentioned above.
Goddammit I hate this one. A TON OF LITTLE PROGRAMS ARE DISTRIBUTED AS EITHER SOURCE OR ONE OF THE FORMATS LISTED.
Then install to /usr/local or /opt, and don't mess up the operating system your distributors have worked so hard to perfect.
Better yet, package the files properly yourself. It's not rocket science. Or you can submit a project recommendation to your distro's wishlist.
Just because the program isn't available for your particular distribution, don't screw up your system by installing packages that weren't made for it. It's just asking for trouble.
Do not carry the Windows philosophy over to Linux. If you do not understand the fundamental concepts behind package managers, do not mess around with them. Most of them do their jobs extremely well unless you decide to try and do their jobs for them.
So the point is, we need something equivalent to "Add/Remove Programs" that just *works* on all linux distros.
Wrong. That's something a Windows user just beginning a migration to Linux would say.
Distributions are different. Full stop. If they were all the same, there would be no point in having several. But assume there is a point in having different distributions, because if there isn't then your sentence is wrong in an even more fundamental manner.
How are distributions different? They're different in where they place their files, how they split up the same source tarball into different packages, what package manager they use, etc. If you take the binary files from a Red Hat .rpm and try using them on Debian, you're on your own, even if your architectures are the same, because the configuration will most likely not be the same as what a Debian maintainer would use.
Long story short: do not install a .rpm on Debian, don't install a SuSE .rpm on Fedora, don't install a .deb on SuSE, etc. If you do, you're just asking for trouble.
There are already Linux equivalents of Windows' Add/Remove Programs (minus the godawful interface). Synaptic, YaST, etc.
In short, there is no problem here. Work on more important stuff (such as integration on the desktop).
During daylight savings you have to use 1, not 12.
But that's ignoring a bigger problem: the "halfway" point is really halfway when the sun's path is 45 degrees away from you. During summertime you'll have to bias yourself to look closer to the 12, and during wintertime farther from it.
In fact, the only place where the rule is universally true is the North Pole, since every direction is South. But there's another problem: the trick will only work for half the year.
I was excited to look at these benchmarks because I know that Linux does 64 bits and I really, really want to see what kind of difference there is before buying an Athlon64. But there is no 64-bit testing. What is the point of running on Linux if you don't take advantage of what Linux offers?
Quality Control does not necessarily mean "user interface". You can make a very high quality cell phone that still sucks. Nokia has an entire factory full of them.
I used to have a Nokia 3390 and it even though it wasn't perfect, it wasn't far off. One big button does what you expect to do all the time. So you just press the big button a few times, and it's all logical.
But that phone died, so I got a Siemens M55. Warning to anybody shopping for a phone: do not get a Siemens M55. Of all the useless features, this one has useless blinking lights on the sides of the phone -- and they constantly blink by default! The confirmation dialogs don't have a consistent placement for Yes/No and the games take ages to load. But that's not what turned me off this phone -- the fact of the matter is, I can't tell when people are calling me! The vibrator is weak and the ringer is quiet. If I'm walking down the street and somebody calls, I'll miss the call.
I'm returning it for a Sony Ericcson T300. I hope it's not just the same story all over again, but I suspect it will be.
All I want is a small, usable phone which also does SMS.
Who comes up with these numbers? Obviously somebody severely misinformed or dishonest. Who would benefit?
The truth is, it's practically free to secure against any new virus. Once your system is reasonably secure you'll only incur maintenance costs, almost never incident costs.
To NVidia's credit, they seem to be somewhat serious about supporting Linux in a somewhat timely manner.
It feels to me like nVidia is a big corporate giant, bending down to let the OSS community eat out of its grubby paws whatever scraps it leaves them.
There's not "always" something to complain about. There's only one thing to complain about, and that's that nVidia does not co-operate with the Linux community. If it did, nVidia drivers would be in the kernel tree and it'd be twice as easy for all of us to get our cards working on a do-it-yourself kernel. Not to mention, we'd have fewer bugs and probably faster framerates.
Linux would benefit. nVidia would benefit.
Drivers are commodities, there's really no good reason to keep them closed. ESR has written plenty of convincing material on the matter, so I won't re-hash it.
Does anyone know how to install this new version?
Well, not sure how well it would work on stable, but on unstable if you roll your own kernel using make-kpkg it's easy:
sudo apt-get install nvidia-kernel-sourcecd
tar zxvf nvidia-kernel-source.tar.gz
cd linux-2.6.1
fakeroot make-kpkg modules sudo dpkg -i
sudo apt-get install nvidia-glx
Back in my day, we had nothing but some floating molecules
You were lucky to have molecules! We hadn't even heard of quarks yet! Molecules? Hmph.
[Student teacher trust is] being violated so regulary by students cheating that teachers wonder if recent degreed graduates really learned anything.
As a student of McGill, I can say quite confidently that no student in arts can graduate without learning a hell of a lot, cheating or not. As for science and engineering, teachers accept as a matter of course that students will collaborate, so assignments are worth very little (i.e., 10%-20% of the grade) and the rest is midterms and final exams (ignoring the issue of cheating on final exams).
I mean, really. A student can't plagiarize a 20-page paper without getting caught, right? But for a 5-page paper it might be possible. The student would either: a) submit something downloaded off the Internet, which is very easy to detect, or b) modify something off the Internet to make it hard to recognize -- thereby learning about the subject, because it would take a lot of work not to get caught.
So I think any student who plagiarizes without getting caught is doing a hell of a lot of work anyway, and is learning. Which is good, because you can't quite plagiarize on midterms and final exams, which account for most of the marks in most courses (even arts).
CVS, like most version control systems, has the concept of "branches"
This is just one example of the article's condescension. Duh. CVS, like most version control systems, also sucks enormously. The Linux kernel development process has the concept of "branches" too (and has had it long before Bitkeeper came around).
In fact, the entire article can basically be summed up in one sentence understandable to the entire target audience:
"BSD is an entire distribution maintained by the same people who hack at its base."
We all know that Netscape 4 is an awful, crashy, buggy, standards-breaking piece of crap that set the Internet back years.
Yes, it was an awful, crashy, buggy, standards-breaking piece of crap; but it certainly wasn't the awful, crashy, buggy, standards-breaking piece of crap which set the Internet back years -- and is still doing so.
I'm sure I can't be the only one thanking Microsoft for the free Linux advertisement. These studies can only be good for Linux, I think.
The GPL states that you only have to provide the sources along with the distibution. The distibution is in this case embedded in their player, so the only thing they would have to answer to is a demand from a verified owner of the player.
Or a demand from anybody who downloaded the ROM from their website, such as the MPlayer developers.
Dell threw in a free DVD+RW upgrade on my Inspiron 600m laptop that I bought this June.
Whoever picked the managers, and supervised the managers, is as much to blame as the damn foam chunk.
I'd say they're entirely to blame. It's happened before, too, with the Challenger disaster. Management was told that Challenger would explode, but they launched anyway.
With the Columbia disaster, critical safety measures were being overlooked. Concerned engineers asked three times for satellite photographs of Columbia, and never got them. Management knew about the foam problem for ages but never helped fix it.
The article basically is correct in stating that passionate dissagreements fork projects. The doubling up of energies on very similar projects (like Gnome and KDE) work against open source.
GNOME and KDE are not forks of the same project. But I'd agree that the general rule of thumb with forks still applies to them:
Forks are healthy.
Let's back it up with examples: egcs/gcc; XFree86/Xouvert/freedesktop.org; samba/samba-tng; apache 1.x/2.x; galeon/epiphany... the list goes on.
I think the examples show that forking is very healthy for open-source projects. In fact, I can't even think of a counterexample. Is there a single case in which a fork resulted in lower-quality code or stagnant development on both branches?
"I am terrified of the following scenario, which is extremely probable to happen soon," responded Jan Rychter in a Wednesday post. "2.4 is being moved into 'pure maintenance' mode and people are being encouraged to move to 2.6. While people slowly start using 2.6, Linus starts 2.7 and all kernel developers move on to the really cool and fashionable things. 2.6 bug reports receive little attention, as it's much cooler to work on new features than fix bugs."
What insight! I can't imagine how anybody could see that far into the future as Jan Rychter.
So many people use the "cooler to work on new features" argument without evidence. Since when has Linux favoured new features over fewer bugs? Linux 2.4 bugs receive plenty of attention, thank you very much -- so do Linux 2.2 bugs and 2.0 bugs.
What kind of person spends that much time trying to find exploits in operating system kernels?
The kernel developers, i.e., Andrew Morton. Good for him, too.
There *was* a patch before the Debian systems were compromised. Hopefully in the future these things will be given more attention before they blow up.
Heh. My apartment has two 25-amp circuits -- appliances included.
My three computers are on the same circuit as the fridge, microwave, toaster and popcorn popper. If we run any three (or sometimes even two) of these appliances at the same time, the fuse blows.
Blown fuses are especially common during the summer, when I have to add on an air conditioner.
And Microsoft's Passport thing? Isn't it meant to include that functionality as well?
No, read the story again. It distinctly says, "a convenient and secure system" (emphasis added).
I couldn't help but notice, this guy put "WebCT" and "quality software" in the same sentence. Just because the software (online teaching suppliment) sells a lot doesn't make it good. I've seen a hell of a lot of bad free software, but never as bad as WebCT.
I'm having to reboot back to 2.4 for the Cisco vpn driver
Though MPPE (weak Microsoft encryption) is not available for 2.6 *yet*, you might want to look into pptpclient. The web page tells you exactly how to set up ppp and everything. It'll connect to pretty much any Windows-ish (i.e., Cisco) VPN, and it's free. MPPE support needs a kernel patch for 2.4 and is on its way for 2.6.
Advantages: GPL, better documentation, the primary developer is very helpful on the mailing list. Disadvantage: maybe harder to set up (not sure, never tried the Cisco one).
I'll bet 2.6 will have MPPE support built in before Cisco releases a 2.6 version of its VPN client.