Nevertheless, like the *theory* of evolution, it is well established by quite a bit of math and experimental evidence. It isn't perfect by any means; physicists have been trying to unify it with other theories like relativity for quite some time, and haven't succeeded yet. But you don't just toss it out the door on a whim. Anything that claims to "disprove" quantum theory is going to be heavily criticized before it has a chance of being accepted.
I think you are arguing a niche case. Software that is not included in the repositories of major distributions falls into one of two categories in the vast majority of cases:
1) software is in alpha stages, under heavy development, the working code is in CVS, or the code is otherwise not intended for general consumption
2) software like, for example, tnimage, uses an old proprietary toolkit making it hard to package on a modern system, and/or software like ygl that just isn't used by more than a handful of people and isn't on the radar of the packagers
The vast majority of the software I use (from office suites, to chemical drawing software, to SNES emulators, to compilers, to obscure libraries for doing different kinds of data analysis) are all available in the Debian repository. I just browse/search using the graphical packaging tool (Oh look, there is even an entry under the Administration menu in Gnome) and the software is installed with no problems. I think the reason you find that most Linux users don't care about this is because the packaging system works great >95% of the time (100% of the time if you use only stock applications). There are problems that various people have acknowledged, but it isn't a high priority issue when it works most of the time and there are a lot of other outstanding issues to worry about.
I agree that in cases where there are no packages, software installation is a bit of a pain. As a couple of people have already said, Autopackage is trying to solve that problem. But, as with any tool, the developers have to actually use it if it is going to be an effective solution. If the software falls into one of the above cases, chances are it isn't going to happen, and that just isn't the fault of Linux. The tools are there, but the projects aren't using them. It's just like if a couple of hackers put together a neat Windows program, didn't bother making an MSI installer, and said "Well, just compile this little program, copy a bunch of files here, here, and here, add this registry key...." You wouldn't say that is the fault of Windows. That is a few people not using the tools available.
Oh, and as for cross-distro compatibility, the main Ubuntu guy has a blurb about binary compatibility in his FAQ. Basically, his stance on the matter is that Open Source is about source code. There have been several attempts to get binary compatibility between distros, but it is a difficult process and in the end prevents distros from doing a lot of things that might make them better for the particular groups they serve. If you look at the Linux kernel, they break the ABI all the time, which is a pain for driver developers, but it allows them to fix things that need fixing, optimize things that need optimizing, and to introduce new features. So they focus on maintaining the API and not worrying about the ABI. With a standardized API, drivers can always be recompiled.
Statically compiling applications is one solution that has been used, as in the cases mentioned with Opera etc.... But that has its own problems. In the end distributions have decided to stick with dealing with dependency hell problems rather than static compiling problems.
Public involvement, yes. But the problem is public involvement is often simply just a NIMBY response like you say. The thing is there is this nuclear waste issue. It is a problem that needs to be solved. The NRC is proposing ways to deal with the problem. It would be nice if the public actually tried to be constructive, and instead of just saying "no way" actually offered a reasonable alternative. But the nature of modern politics is to simply be confrontational and devisive, not to actually solve problems. So this particular problem has been around for 50 yrs, and it doesn't look like it is going to be solved in the next 20.
You just need to grab the madwifi drivers. I'm kind of surprised Suse doesn't already have them because Debian/Ubuntu and Gentoo have had them for at least a year and a half. Of course the problem might not be the driver, but rather the encryption. No linux distribution I have tried yet handles WPA out of the box. You still have to mess around with wpa_supplicant which is kind of a pain. Not sure why this hasn't been integrated into the network control panels yet like everything else.
"Users were confused because they expected the username and password fields to be on the same screen. Suggestion: change the login manager to put the username and password fields on the same screen."
"Users were confused because double-clicking the clock applet didn't bring up time/date setting. Suggestion: make the applet behave more like it does in Windows."
"Users were confused about the whole 'enter the root password' thing. Suggestion: make it so entering the root password isn't necessary."
What kind of crap is this? This isn't usability testing. It is an excercise in how to make the linux desktop behave exactly like Windows, security problems and all.
(just because GWB is a punishment-oriented authoritarian doesn't mean those are the values of the people who are registered republicans)
Well, since it is the registered republicans that voted for him, I would have to conclude that those are their values. If a politician isn't representing your interests, you shouldn't vote for him, right? Unless you want to argue that our government isn't actually repres...err, dum de doo bah dum.
Well, there is and should be a certain amount of "let the buyer beware." If I setup a webserver, I am going to pick Apache over IIS because I think the quality is much better. Likewise when I pick Postgresql over MSSQL. However, if a company markets a product to do something, it needs to make a reasonable effort to eliminate defects. If I'm driving down the road at 60 mph and the wheels fall off my car, you better damn be sure I will and should be able to sue for it. Likewise, if I'm using a financial application for my business and I lose a million dollars in a transaction due to a floating point error (I believe this actually happened a few years ago, don't remember the details), I better be able to get some recompense. Or if my tax management software results in me getting a huge penalty because it didn't calculate my deductions right, etc....
I guess it comes down to this: if I pay for the software to perform a task as marketed, and the software fails resulting in excessive injury to me, I should have some recourse. If it is just a matter of a BSOD while I'm typing my thesis, I'll be hugely inconvenienced, but nothing more. In the former case, "let the buyer beware" is not sufficient because I have already been injured through no ignorance or fault of my own (i.e: I researched the product and the company and had reason to believe that the software would perform as marketed). If the company isn't charging enough to make a guarantee--which is hard to believe given typical profit margins in software--then that is the company's fault. And no, I shouldn't have to pay 1000 times more for a given piece of software just to get a reasonable degree of quality. Notice how I can buy the cheapest and crapiest car available on the market and still be able to sue the car company if a defect in their product results in property damage or an injury.
Binary compatibility was never more than a nasty hack, fudged in for the benefit of those who want to lock up the source code of their software. These people are pure evil. By not sharing their code with you, they are just one very tiny step removed from stealing from you.
You were making a good point until here. This just makes you sound like a fanatic. I personally prefer open source software and will lobby for and support vendors who distribute open source software. But proprietray vendors are *not* "evil." Their business model is selling the program, so they have to keep the source secret. I happen to think that is a lousy (but profitable for the time being) business model, but that doesn't make them evil.
The problem with binary incompatibility is the need for either the software vendor to compile for every possible distribution, or the distribution vendor to compile every possible piece of software (no, you can't expect the end user to do it). That is not an ideal situation. LSB may not be the solution, but some sort of standards are necessary. If you *need* to break the standard for some reason fine, but you shouldn't break it "just 'cause." To use your own example, if every appliance vendor used their own special plug, you would have to have your house modified before you could buy it. Don't you think it is a little better for them to be following a standard?
I've been using Linux for many years, and the problem of obtaining software packages drives me to the end of my nerves.
Hmmm, what packages are you trying to install? I usually only have that problem if the software is still in CVS or something. But even then, Gentoo still usually packages it. I do have problems when I use Fedora. Not sure why they don't have a lot of packages, but it is hard to find stuff sometimes. With Debian, though, almost never a problem. In any case, Autopackage is supposed to solve this problem. It actually looks pretty neat, but I am waiting for local packaging system integration. Then it will just be a matter of convincing developers to use it. Some projects, like Abiword, already are.
Well, maybe this is revisionist history, but I do seem to remember ReiserFS being the first journalling filesystem. XFS and JFS came soon after, but ReiserFS was in the main kernel first and was supported by mainstream distributions right away. I remember because at the time there was a really big deal about Linux not having a journalling filesystem, and then there was ReiserFS. Feel free to correct me, though, if you can find an old news blurb or something.
I have no doubt that Microsoft code is becoming more secure and less exploitable. However, I am concerned about a couple of other issues that I'm not sure their new security initiatives are addressing:
1) A secure computing environment. That is, not just secure applications, but closed ports, stringent remote access requirements, disabled unnecessary services, privilege separation, and in general the ability to let people run applications on your computer while restricting access to computer resources that they shouldn't access (listening on privileged ports, for example).
2) A secure development environment. That is, providing the tools and documentation needed to allow third-party developers to write secure applications. I still run into tons of applications that just can't be run unless you are an administrator.
Is Microsoft doing anything about these, or is it just combing its code for buffer overflows?
And it's probably because you run your own business and studied economics that you're blinded to other nonstandard possibilities.
That and the economics a person learns in high school is just that, high school economics. The whole "I'll give you this service in exchange for that service" is a very basic model. Real economics is a lot more complicated, and current models are constantly being debated and revised (kind of like science, because it is a social science) as more is learned about what motivates people to create things or exchange with other people. The posit that people are completely self-interested and only motivated by opportunities to better their personal well-being (the basis for Milton Friedman economics) is heavily criticized in many academic circles, and old views about free market capitalism are starting to change.
Point taken. But a computer has always been a multi-purpose device, whereas a phone or a stereo has always been a standalone appliance. I don't think mixing a phone and a stereo together is all that great, mostly because I would rather get something like the iRiver, which supports Ogg and video, and use it without depleting the battery on my phone which gets used up fast enough as it is.
It is kind of like Mozilla Suite vs. Mozilla Firefox. Some people like the entire suite because they want to use the addressbook, calendar, email, composer, browser, chat, kitchen sink all at once in one giant application. I, and a lot of people however, prefer to use just the browser because it is excellent. I don't chat that much, so I don't want an irc client super-integrated into my web browser.
Ok, maybe not directly contradictory. However, the poster was expressing an idealogical viewpoint: people should not let liberty be taken from them. This requires effort and political involvement. You can't just sit comfortably on your couch, watching tv and eating Doritos. You actually have to pay attention to what the government and other people around you are doing, and react vocally to things you see as threatening to your liberty.
On the other hand, though, he says people with a different idealogical viewpoint--advocating for open source drivers--are just getting in the way of a practical solution that is being handed to them. In other words, he is saying sit back, be comfortable, take what is given to you because it is good, and don't try to to get something better because good is good enough. To me that is a contradictory set of opinions even if they are not directly contradictory from the given statements.
On the practicality point: just as it is more practical to not give up liberty in the first place, it is also more practical to have an open source driver that can not only be distributed with Xorg but be properly integrated into the kernel. I think the difference in opinion stems from the one camp that is happy that their computer works despite any hoops they have to jump through, and the other camp that wants their computer to work better and doesn't want to jump through any hoops. It is not really a practical vs. ideological debate even though it is framed that way.
Gtk just added Cairo support in this release, and Cairo doesn't quite yet have a fully functioning Glitz backend. I would keep an eye out for more eye candy in the next few releases because the infrastructure is there now, it just needs to be fleshed out.
Showing the Desktop in two places is also inconsistent with the spatialness of Gnome.
I sort of see your point with this, but I don't really see it as a usability problem. I think most users will realize that the Desktop folders contains the items that show up on their desktop. I don't think hiding the Desktop folder aids usability; it only irritates system administrators. Although you are right, it does violate the spatial metaphor.
Candidates for default hiding would be:/etc,/proc,/usr/,/bin,/lib,/sbin,/dev,/root
No, and fsck no. The correct way of hiding system complexity is the way it is already done. The user is rooted in their home directory. The Home bookmark or whatever you want to call it is prominently placed in the file choose dialog, the desktop, the places menu, and the computer window. All views of the filesystem default to the home directory, and the only way to get to the rest of the computer is by going "Up" from the home directory (which is not an obvious action if you are navigating with the location bar), or clicking on the Filesystem bookmark. The completely borked up Windows way of doing it--hiding anything system related unless you click the "show contents" button--is not only extremely irritating, but it doesn't work.
Several lawsuits, as technology used in writing those drivers is patented
It depends. I'm sure some technologies may be patented, but what do you need in a driver? And OpenGL implementation? Well a free one, Mesa, has existed for years without worrying about patents. A DirectX implementation? No, that's all done by Microsoft. Interfacing with the hardware. Now that's probably patented, but hopefully by themselves. No, I don't think the patent argument holds water unless they chose to use somebody else's code instead of writing their own (in which case it still probably wouldn't be a patent issue, but rather a licensing issue).
the drivers would still be a 'trade' secret,
Most open driver advocates, including myself, consider that statement complete bullshit. The way a driver talks to the actual hardware does nothing to reveal trade secrets any more than publishing schematics or even just looking at the friggen card. Anybody with an electrical engineering degree and some time can figure out how the card works simply by possessing it. That in and of itself does not allow them to come up with a better competing card.
They who would give up an essential liberty for temporary security, deserve neither liberty or security.
is a political statement, and the grandparent said in a previous paragraph that practical result is more important than idealogical view (i.e: contradictory opinions about the importance of political involvement). The parent did not say anything about "liberty" in an Nvidia product, nor do I think he implied it.
I do a lot of fairly advanced stuff, and I wish to do it without ever using a command line, or editing a configuration file by hand.
Ok, fine. You are right, this is not a reasonable expectation on linux. It is not even a reasonable expectation on OS X because if you are trying to, say, setup the Apache web server, you have to edit its config by hand. The fact that you can get away with it on Windows is not something I would consider good. I think a lot of dialogs with a lot of unexplained advanced options organized in a way that makes sense to Microsoft, but not to me, makes the user interface cluttered, makes it hard to find options if you know what you are looking for, and encourages people to randomly click on things without knowing what they are doing. And don't even try to claim that the management console is easy to use because it certainly isn't.
But my posts, and I believe that of the original poster, were not aimed at "power users." I agree, you can't do advanced stuff without the command line in linux, but I also don't consider this a bad thing. And by "advanced stuff" I mean setting up servers, installing third-party kernel modules, and compiling software. Things like installing software, changing network settings (wifi excepted, this is being worked on), and setting up printers can all be done from the gui. So if you are a "basic user" who needs to occasionally install some software or wants to access a usb stick, this can be done just as easily on linux as on Windows, provided you are willing to learn how to do it (because it is different than on Windows).
Actually I believe there are ext3 drivers for Windows, and there certainly must be something for OS X. So there you go, a journalling file system supported by all three operatings systems.
Well, the point I was trying to make is that it is completely unintuitive to have to find this option and uncheck it before revealing the security tab. If I just want to look at and modify the file permissions, and I'm not aware of this quirk, it can take a long time to figure this out.
If you want to get basic stuff done, it is just as easy on linux as Windows. Just click the Openoffice icon instead of the Microsoft Office icon. And I think that's the point of the whole thread. Linux isn't hard, Windows isn't easy, it is just a matter of getting the feel for the os. If find Windows quite frustrating because I don't know all of the quirks, and the quirks I do know about I think are stupid. I'm sure you can say the same thing about linux.
And just for the record, the only time I ever come close to looking at a manual in linux is when I am at the command line and I do a tar --help to get a list of flags for tar or something. Even if you have to edit a.conf file by hand, which you don't if you are doing basic things, they are often well-documented and self-explanatory. I like to do a lot of things at the command line because I think it is faster and more efficient, but linux has a pretty decent gui these days.
Slashdot Mirror
Nevertheless, like the *theory* of evolution, it is well established by quite a bit of math and experimental evidence. It isn't perfect by any means; physicists have been trying to unify it with other theories like relativity for quite some time, and haven't succeeded yet. But you don't just toss it out the door on a whim. Anything that claims to "disprove" quantum theory is going to be heavily criticized before it has a chance of being accepted.
I think you are arguing a niche case. Software that is not included in the repositories of major distributions falls into one of two categories in the vast majority of cases:
1) software is in alpha stages, under heavy development, the working code is in CVS, or the code is otherwise not intended for general consumption
2) software like, for example, tnimage, uses an old proprietary toolkit making it hard to package on a modern system, and/or software like ygl that just isn't used by more than a handful of people and isn't on the radar of the packagers
The vast majority of the software I use (from office suites, to chemical drawing software, to SNES emulators, to compilers, to obscure libraries for doing different kinds of data analysis) are all available in the Debian repository. I just browse/search using the graphical packaging tool (Oh look, there is even an entry under the Administration menu in Gnome) and the software is installed with no problems. I think the reason you find that most Linux users don't care about this is because the packaging system works great >95% of the time (100% of the time if you use only stock applications). There are problems that various people have acknowledged, but it isn't a high priority issue when it works most of the time and there are a lot of other outstanding issues to worry about.
I agree that in cases where there are no packages, software installation is a bit of a pain. As a couple of people have already said, Autopackage is trying to solve that problem. But, as with any tool, the developers have to actually use it if it is going to be an effective solution. If the software falls into one of the above cases, chances are it isn't going to happen, and that just isn't the fault of Linux. The tools are there, but the projects aren't using them. It's just like if a couple of hackers put together a neat Windows program, didn't bother making an MSI installer, and said "Well, just compile this little program, copy a bunch of files here, here, and here, add this registry key...." You wouldn't say that is the fault of Windows. That is a few people not using the tools available.
Oh, and as for cross-distro compatibility, the main Ubuntu guy has a blurb about binary compatibility in his FAQ. Basically, his stance on the matter is that Open Source is about source code. There have been several attempts to get binary compatibility between distros, but it is a difficult process and in the end prevents distros from doing a lot of things that might make them better for the particular groups they serve. If you look at the Linux kernel, they break the ABI all the time, which is a pain for driver developers, but it allows them to fix things that need fixing, optimize things that need optimizing, and to introduce new features. So they focus on maintaining the API and not worrying about the ABI. With a standardized API, drivers can always be recompiled.
Statically compiling applications is one solution that has been used, as in the cases mentioned with Opera etc.... But that has its own problems. In the end distributions have decided to stick with dealing with dependency hell problems rather than static compiling problems.
Public involvement, yes. But the problem is public involvement is often simply just a NIMBY response like you say. The thing is there is this nuclear waste issue. It is a problem that needs to be solved. The NRC is proposing ways to deal with the problem. It would be nice if the public actually tried to be constructive, and instead of just saying "no way" actually offered a reasonable alternative. But the nature of modern politics is to simply be confrontational and devisive, not to actually solve problems. So this particular problem has been around for 50 yrs, and it doesn't look like it is going to be solved in the next 20.
You just need to grab the madwifi drivers. I'm kind of surprised Suse doesn't already have them because Debian/Ubuntu and Gentoo have had them for at least a year and a half. Of course the problem might not be the driver, but rather the encryption. No linux distribution I have tried yet handles WPA out of the box. You still have to mess around with wpa_supplicant which is kind of a pain. Not sure why this hasn't been integrated into the network control panels yet like everything else.
Yeah, and did you read any of their reports.
"Users were confused because they expected the username and password fields to be on the same screen. Suggestion: change the login manager to put the username and password fields on the same screen."
"Users were confused because double-clicking the clock applet didn't bring up time/date setting. Suggestion: make the applet behave more like it does in Windows."
"Users were confused about the whole 'enter the root password' thing. Suggestion: make it so entering the root password isn't necessary."
What kind of crap is this? This isn't usability testing. It is an excercise in how to make the linux desktop behave exactly like Windows, security problems and all.
(just because GWB is a punishment-oriented authoritarian doesn't mean those are the values of the people who are registered republicans)
Well, since it is the registered republicans that voted for him, I would have to conclude that those are their values. If a politician isn't representing your interests, you shouldn't vote for him, right? Unless you want to argue that our government isn't actually repres...err, dum de doo bah dum.
Yeah, and linux has had it longer than that. What's your point?
Well, there is and should be a certain amount of "let the buyer beware." If I setup a webserver, I am going to pick Apache over IIS because I think the quality is much better. Likewise when I pick Postgresql over MSSQL. However, if a company markets a product to do something, it needs to make a reasonable effort to eliminate defects. If I'm driving down the road at 60 mph and the wheels fall off my car, you better damn be sure I will and should be able to sue for it. Likewise, if I'm using a financial application for my business and I lose a million dollars in a transaction due to a floating point error (I believe this actually happened a few years ago, don't remember the details), I better be able to get some recompense. Or if my tax management software results in me getting a huge penalty because it didn't calculate my deductions right, etc....
I guess it comes down to this: if I pay for the software to perform a task as marketed, and the software fails resulting in excessive injury to me, I should have some recourse. If it is just a matter of a BSOD while I'm typing my thesis, I'll be hugely inconvenienced, but nothing more. In the former case, "let the buyer beware" is not sufficient because I have already been injured through no ignorance or fault of my own (i.e: I researched the product and the company and had reason to believe that the software would perform as marketed). If the company isn't charging enough to make a guarantee--which is hard to believe given typical profit margins in software--then that is the company's fault. And no, I shouldn't have to pay 1000 times more for a given piece of software just to get a reasonable degree of quality. Notice how I can buy the cheapest and crapiest car available on the market and still be able to sue the car company if a defect in their product results in property damage or an injury.
Binary compatibility was never more than a nasty hack, fudged in for the benefit of those who want to lock up the source code of their software. These people are pure evil. By not sharing their code with you, they are just one very tiny step removed from stealing from you.
You were making a good point until here. This just makes you sound like a fanatic. I personally prefer open source software and will lobby for and support vendors who distribute open source software. But proprietray vendors are *not* "evil." Their business model is selling the program, so they have to keep the source secret. I happen to think that is a lousy (but profitable for the time being) business model, but that doesn't make them evil.
The problem with binary incompatibility is the need for either the software vendor to compile for every possible distribution, or the distribution vendor to compile every possible piece of software (no, you can't expect the end user to do it). That is not an ideal situation. LSB may not be the solution, but some sort of standards are necessary. If you *need* to break the standard for some reason fine, but you shouldn't break it "just 'cause." To use your own example, if every appliance vendor used their own special plug, you would have to have your house modified before you could buy it. Don't you think it is a little better for them to be following a standard?
I've been using Linux for many years, and the problem of obtaining software packages drives me to the end of my nerves.
Hmmm, what packages are you trying to install? I usually only have that problem if the software is still in CVS or something. But even then, Gentoo still usually packages it. I do have problems when I use Fedora. Not sure why they don't have a lot of packages, but it is hard to find stuff sometimes. With Debian, though, almost never a problem. In any case, Autopackage is supposed to solve this problem. It actually looks pretty neat, but I am waiting for local packaging system integration. Then it will just be a matter of convincing developers to use it. Some projects, like Abiword, already are.
Nasty stuff. Toxic, and readily absorbed through the skin.
Ummm...so kind of like, gasoline? Or how about lead and sulfuric acid?
Well, maybe this is revisionist history, but I do seem to remember ReiserFS being the first journalling filesystem. XFS and JFS came soon after, but ReiserFS was in the main kernel first and was supported by mainstream distributions right away. I remember because at the time there was a really big deal about Linux not having a journalling filesystem, and then there was ReiserFS. Feel free to correct me, though, if you can find an old news blurb or something.
I have no doubt that Microsoft code is becoming more secure and less exploitable. However, I am concerned about a couple of other issues that I'm not sure their new security initiatives are addressing:
1) A secure computing environment. That is, not just secure applications, but closed ports, stringent remote access requirements, disabled unnecessary services, privilege separation, and in general the ability to let people run applications on your computer while restricting access to computer resources that they shouldn't access (listening on privileged ports, for example).
2) A secure development environment. That is, providing the tools and documentation needed to allow third-party developers to write secure applications. I still run into tons of applications that just can't be run unless you are an administrator.
Is Microsoft doing anything about these, or is it just combing its code for buffer overflows?
And it's probably because you run your own business and studied economics that you're blinded to other nonstandard possibilities.
That and the economics a person learns in high school is just that, high school economics. The whole "I'll give you this service in exchange for that service" is a very basic model. Real economics is a lot more complicated, and current models are constantly being debated and revised (kind of like science, because it is a social science) as more is learned about what motivates people to create things or exchange with other people. The posit that people are completely self-interested and only motivated by opportunities to better their personal well-being (the basis for Milton Friedman economics) is heavily criticized in many academic circles, and old views about free market capitalism are starting to change.
Point taken. But a computer has always been a multi-purpose device, whereas a phone or a stereo has always been a standalone appliance. I don't think mixing a phone and a stereo together is all that great, mostly because I would rather get something like the iRiver, which supports Ogg and video, and use it without depleting the battery on my phone which gets used up fast enough as it is.
It is kind of like Mozilla Suite vs. Mozilla Firefox. Some people like the entire suite because they want to use the addressbook, calendar, email, composer, browser, chat, kitchen sink all at once in one giant application. I, and a lot of people however, prefer to use just the browser because it is excellent. I don't chat that much, so I don't want an irc client super-integrated into my web browser.
Just telnet your mail host port 25 and if you are stuck type help.
Yes, and a lot of ISPs these days require you to authenticate to connect to port 25 for that exact reason. So it would still be pretty hard.
Ok, maybe not directly contradictory. However, the poster was expressing an idealogical viewpoint: people should not let liberty be taken from them. This requires effort and political involvement. You can't just sit comfortably on your couch, watching tv and eating Doritos. You actually have to pay attention to what the government and other people around you are doing, and react vocally to things you see as threatening to your liberty.
On the other hand, though, he says people with a different idealogical viewpoint--advocating for open source drivers--are just getting in the way of a practical solution that is being handed to them. In other words, he is saying sit back, be comfortable, take what is given to you because it is good, and don't try to to get something better because good is good enough. To me that is a contradictory set of opinions even if they are not directly contradictory from the given statements.
On the practicality point: just as it is more practical to not give up liberty in the first place, it is also more practical to have an open source driver that can not only be distributed with Xorg but be properly integrated into the kernel. I think the difference in opinion stems from the one camp that is happy that their computer works despite any hoops they have to jump through, and the other camp that wants their computer to work better and doesn't want to jump through any hoops. It is not really a practical vs. ideological debate even though it is framed that way.
Gtk just added Cairo support in this release, and Cairo doesn't quite yet have a fully functioning Glitz backend. I would keep an eye out for more eye candy in the next few releases because the infrastructure is there now, it just needs to be fleshed out.
Showing the Desktop in two places is also inconsistent with the spatialness of Gnome.
/etc, /proc, /usr/, /bin, /lib, /sbin, /dev, /root
I sort of see your point with this, but I don't really see it as a usability problem. I think most users will realize that the Desktop folders contains the items that show up on their desktop. I don't think hiding the Desktop folder aids usability; it only irritates system administrators. Although you are right, it does violate the spatial metaphor.
Candidates for default hiding would be:
No, and fsck no. The correct way of hiding system complexity is the way it is already done. The user is rooted in their home directory. The Home bookmark or whatever you want to call it is prominently placed in the file choose dialog, the desktop, the places menu, and the computer window. All views of the filesystem default to the home directory, and the only way to get to the rest of the computer is by going "Up" from the home directory (which is not an obvious action if you are navigating with the location bar), or clicking on the Filesystem bookmark. The completely borked up Windows way of doing it--hiding anything system related unless you click the "show contents" button--is not only extremely irritating, but it doesn't work.
No shit. What's your point?
Several lawsuits, as technology used in writing those drivers is patented
It depends. I'm sure some technologies may be patented, but what do you need in a driver? And OpenGL implementation? Well a free one, Mesa, has existed for years without worrying about patents. A DirectX implementation? No, that's all done by Microsoft. Interfacing with the hardware. Now that's probably patented, but hopefully by themselves. No, I don't think the patent argument holds water unless they chose to use somebody else's code instead of writing their own (in which case it still probably wouldn't be a patent issue, but rather a licensing issue).
the drivers would still be a 'trade' secret,
Most open driver advocates, including myself, consider that statement complete bullshit. The way a driver talks to the actual hardware does nothing to reveal trade secrets any more than publishing schematics or even just looking at the friggen card. Anybody with an electrical engineering degree and some time can figure out how the card works simply by possessing it. That in and of itself does not allow them to come up with a better competing card.
I think the point he was making is that
They who would give up an essential liberty for temporary security, deserve neither liberty or security.
is a political statement, and the grandparent said in a previous paragraph that practical result is more important than idealogical view (i.e: contradictory opinions about the importance of political involvement). The parent did not say anything about "liberty" in an Nvidia product, nor do I think he implied it.
I do a lot of fairly advanced stuff, and I wish to do it without ever using a command line, or editing a configuration file by hand.
Ok, fine. You are right, this is not a reasonable expectation on linux. It is not even a reasonable expectation on OS X because if you are trying to, say, setup the Apache web server, you have to edit its config by hand. The fact that you can get away with it on Windows is not something I would consider good. I think a lot of dialogs with a lot of unexplained advanced options organized in a way that makes sense to Microsoft, but not to me, makes the user interface cluttered, makes it hard to find options if you know what you are looking for, and encourages people to randomly click on things without knowing what they are doing. And don't even try to claim that the management console is easy to use because it certainly isn't.
But my posts, and I believe that of the original poster, were not aimed at "power users." I agree, you can't do advanced stuff without the command line in linux, but I also don't consider this a bad thing. And by "advanced stuff" I mean setting up servers, installing third-party kernel modules, and compiling software. Things like installing software, changing network settings (wifi excepted, this is being worked on), and setting up printers can all be done from the gui. So if you are a "basic user" who needs to occasionally install some software or wants to access a usb stick, this can be done just as easily on linux as on Windows, provided you are willing to learn how to do it (because it is different than on Windows).
Actually I believe there are ext3 drivers for Windows, and there certainly must be something for OS X. So there you go, a journalling file system supported by all three operatings systems.
Well, the point I was trying to make is that it is completely unintuitive to have to find this option and uncheck it before revealing the security tab. If I just want to look at and modify the file permissions, and I'm not aware of this quirk, it can take a long time to figure this out.
.conf file by hand, which you don't if you are doing basic things, they are often well-documented and self-explanatory. I like to do a lot of things at the command line because I think it is faster and more efficient, but linux has a pretty decent gui these days.
If you want to get basic stuff done, it is just as easy on linux as Windows. Just click the Openoffice icon instead of the Microsoft Office icon. And I think that's the point of the whole thread. Linux isn't hard, Windows isn't easy, it is just a matter of getting the feel for the os. If find Windows quite frustrating because I don't know all of the quirks, and the quirks I do know about I think are stupid. I'm sure you can say the same thing about linux.
And just for the record, the only time I ever come close to looking at a manual in linux is when I am at the command line and I do a tar --help to get a list of flags for tar or something. Even if you have to edit a