Also, I want this functionality on all operating systems. Unless I explicitly request deletion, no file should ever be unlinked, deleted, or whatever you call it when I delete it, whether through the command line or the GUI.
The problem with this is, is that it causes a significant reduction in performance.
Ideally, the operating system chose the best possible spot for that file when it got written. Once that file is deleted, that spot will once again be the fastest best possible spot- for at least something. If the operating system skips that spot for a new file, then this new file isn't going to be accessed quite as quickly.
Truly automatic tiered storage solves this problem by splitting the directory services from the storage system- that is, the file's _name_ is no longer tied to the volume that the file happens to live on (and no, this isn't the same thing as symlinks or shortcuts). This allows the decision as to what the best spot for a file is to be deferred until later- and even spanned across multiple volumes!
Unfortunately, such a beast is very difficult- if we make a reduction in our requirements- say that performance isn't very important- or perhaps that we can stop using our computer for a few hours each evening, then it's probably possible. What we need is a new kind of file system that supports either atomic moves between disks, or a filesystem that splits the names from the storage.
A few research projects have been focused on these kinds of changes- but they all tend to break UNIX semantics (Amoeba immediately springs to mind)- and those UNIX semantics are, in-fact, the most widely used and recognized semantics for filesystems anywhere (Even Windows uses them!)-- people who develop a filesystem incapable of supporting them, really need to have a real good reason for breaking everyone's hard work.
While they often do, it hasn't yet been seen as good enough for general purpose stuff.
SFU: Let devs, systems admins port applications SFU: For talking to *nix clients using nfs, ldap and interact with AD
Hah. You've never actually used SFU, have you? NFS support doesn't work unless you authenticate against AD- that means making your AD server your NIS server. SFU's "porting kit", btw, consists of (get this) GCC, CYGWIN, and a number of other free softwares.
RTF Format: As an open spec that predates everything else in the opensource world
Ah, no. TeX predates RTF by about 10 years..Net Runtime as EMCA standard
Check again. The ECMA submitted runtime lacks the entire WinForms interface. That is, if all you want to do is make text-based hello world programs, that's fine, but it's a far cry from usefulness.
Open XML as EMCA standard
It's ECMA. Again, OpenXML is worthless. It's got almost zero marketshare. I wouldn't want to pick up that tar baby either.
Wix install set: For open software to create installers for windows
But only ON windows. It relies on API that aren't public to do so. This negates cross-compilers, and makes farm-work difficult.
IBuySpy Portal(Dotnetnuke is based on)
Again, zero marketshare.
It is easy to blame when you look at only one point of view
I agree.
Sure MS should be releasing docs for smb, cifs, AD, rdp(dont know if they actually control it) etc.
They indeed should be! See below:
should they for free (open) that is debatable....
No it's not. They were so ordered by a US Federal Court. They are convicted criminals, so it's not debatable at all. They continue to break the law, and they continue to hurt Americans and the world at large by continuing to have turd-covered shills pretend that Microsoft is some kind of company. It's not. It's an illegal monopoly. Move on.
I, like many people, like capturing with ethereal rather than tcpdump because it's more convenient. To do this it asks for the root password.
I, like many people, use packet sniffers to analyze network traffic besides the traffic that my workstation might be connected to (which happens to be a switch).
So what? Ethereal has lots of non-privileged user uses.
Microsoft Outlook does not need to be run as an administrator. Palm desktop does not need to be run as an administrator. What is your point here?
In less so a way that Ethereal needs to be run as root. Palm desktop will not run as non-admin without registry hacks and permission effects. Outlook won't run as a regular user unless it has been run on the machine as the administrator.
but like I originally said an administrator can cache credentials - this means you can give a user an icon on a desktop that will run an app as any user including the administrator. Sure it's a pain, but it's due to poorly written software. If I deal with software like this I'll usually look for a suitable alternative first.
So what? My point is that "cached credentials" are not like sudo or su at all. They work on a different principal, and if you allow "one icon" to run as administrator, another application can steal those credentials.
Forgive my being rusty but I don't do a lot of suexec
That's obvious.
Some applications do - ethereal for example will ask for the root password if you are not root when you run it, giving you the option to run it as unpriviliged, but it does ask the user.
No it doesn't. That's a fundimental difference between UNIX versus Windows security models. Ethereal doesn't ask for the root password- userhelper or sudo asks for it- these are setuid programs that masquarade as/usr/bin/ethereal on some systems, so that when you run/usr/bin/ethereal (a program called ethereal, but not actually ethereal) that is a symlink to userhelper which is a setuid tool that becomes root, then asks for your password, and then runs/usr/sbin/ethereal.
Under Windows, ethereal can actually ask you for your password, and having it, it can obtain an authentication token from the system.
This is a very significant difference between the two systems- you say it may not look different, but the security implications are enormous. One way is very easy for administrators and developers to make and be safe, the other is a literal bomb that goes off regularly, and that administrators are at the whim of developers that think it's the job of the administrators.
The original code was proprietary - the folks at Berkeley kinda sorta stole it - didn't they?
No they didn't.
Even if the original Bell lab guys wanted to share, was it theirs to share?
Actually, AT&T shared it because legal obligations prevented them from entering the Operating Systems as a market.
But back in the Bell Labs days, nobody cared. It just wasn't thought about- after all, why would someone buy hardware that they can't do anything with?
That's kind of my question these days- why do people buy hardware that they can't do anything with?
I'm sure I missed some things, but other posters will point them out.
Actually yeah, the BEST way to find out how to get an application to install/run with reduced privileges is to instead search for how to install it in a Domain. The procedures are the same, but in the example of Palm Desktop- one will tell you how to do it and the other will not.
but there are many *NIX apps that expect to have root - ethereal for example.
Uh, no? I generate the dumps with tcpdump and load them into a ethereal running as non-root all the time.... they are usually system admin related...
Which I think is the critical difference. Microsoft Outlook has _zero_ to do with system administration. Palm Desktop has _zero_ to do with system administration.
In Windows you can use the runas command similar to su to give elevated privs to individual apps.
Well, you might say that runas is similar to su or sudo, the difference is that people don't want to, nor should they be using RunAs to open Microsoft Word. It's just stupid.
You can also use a switch to cache credentials (like chown +x root)
Err, did you mean chmod u+s?
On Windows, using cached credentials means the user session can gain Administrator privileges, but on UNIX it means that program can be run with elevated privileges.
On one hand, as long as a program doesn't ask for Administrator privileges, it'll be run in a reduced environment, but on UNIX the programs cannot "ask" to have their privileges elevated.
You're assuming, of course, that no commercial venture would ever challenge MS for web browser domination
No I'm not. I'm assuming that they would fail. I'm assuming this because several companies did challenge MS and lost.
You recall, of course, that Unix was not open source originally, so one could claim that without AT&T there would be no open source.
UNIX most certainly was Open Source originally. See John Lion's commentary on the Sixth Edition UNIX Operating system for more details.
If it weren't for Microsoft, it would've stayed that way.
Nevertheless, I'm not talking about Open Source, I'm talking about Free Software. If UNIX was originally Free Software, it always would be Free Software. The distinction is critical.
You're full of shit because I can still use older MS products.
Software isn't like an automobile or a gallon of milk it doesn't rot. When it comes out of the factory, it doesn't begin to break down and rust, but actually any bugs that are in it are in-fact defects in the software itself.
Security bugs, while a special class, are bugs and are in-fact defects.
So if you could, back in 1998, run Windows 98 directly connected to the Internet comfortably, read your email, and work on your documents, why can't you do it now?
Of course Windows 98 had bugs in it, and of course it still does, but in order to have the productivity that they had in 1998, they can't use the software and the hardware that they had in 1998.
That's the behavior that people are referring to by "forced upgrade".
I don't give one rat's ass if I can see the source code or not
But you do care if I can see the source code. Because I am a developer, and a Free software developer at that.
If Free Software wasn't available, MSIE would still be sitting at 6.0- no PNG and no SVGs- after all, no Free software would mean no Firefox. No competition means no updates- remember before Firefox Microsoft said that MSIE 6 would be the last release!
But we don't have to stop there: If Free software weren't available, you wouldn't have your World-Wide-Web because there wouldn't be a GCC and without a GCC there wouldn't be a NeXT and without a NeXT there wouldn't be a WorldWideWeb.
Source code matters even to non-developers- EVEN when they don't realize that it matters.
Just a week ago I tried to install CVS or subversion to make a Latex document repository, after reading the fucking manual and downloading this Explorer shell extension programs (dont remember their name) I got fed up and gave up, I supposedly made a repository directory and tried to create a new tree or whatever it is called but the darn shit just could not work.
That's very brave- All those millions of people out there who do manage to get CVS to work must simply be smarter than you.
Generally, it's a good idea to ask questions when you don't know the answers.
My employer managed to install TortoiseCVS with no difficulty, and uses it daily to manage his MS-Word documents. If you like, I can give you his email address and you can ask him what he did.
And it is not really that slow and unreliable that for a lot of companies who preffer to run a setup wizzard instead of trying to figure out how to setup CVS or SV.
Again, see, you really should ask questions when you don't know what you're talking about.
CVS can and was setup via "a wizard". Module creation proved to be no difficulty- and I don't think my employer opened the manual at all.
I managed to install CVS via Debian's DPKG system and get a centralized repository setup- all from Debian's wonderful wizard-based system.
So really, while the reason you didn't install CVS or SV is simply that you're not smart enough to, that doesn't need to be a stumbling block. There are lots of people around that are smart enough to set up CVS or SV, and some of them might be quite willing to answer your questions, and help guide you through it.
I can: wanting to get no-cd cracks so as to be able to play all your computer games on your laptop without having to lug 50 CDs around.
Which is not a legitimate use.
Of course that's a legitimate use.
I bet the EULA and/or license for the game forbids you from using such software.
So what? You read this copyrighted post, and to make use of it (understand it), you must agree to these terms: I forbid you from eating grapes!
Does that mean it's illegal to eat grapes? You did read my post, and it's copyrighted, and it has affected you in exactly the same way as any other copyrighted material you otherwise legally receive.
So why are you still legally allowed to eat grapes? Because I can't take away your rights "just by saying so", and neither can software publishers.
but that doesn't mean I want my desktop in Linux to be hobbled because someone thinks it's a cardinal sin to taint the kernel.
The reason for "tainting" messages is so that you understand that you're giving up support. Nobody can help you with a tainted kernel except the people who tainted it (ATI or NVidia in this case).
Why? Because kernel structures have been modified and nobody knows how or why. They don't have source code, and so therefore it's very very difficult to find out where ATI or NVidia screwed it up.
And it's common sense for Pete's sake that onboard video is going to suck hard compared to a full-blown and fully featured dedicated video card
So what? I'm not suggesting on-board video. What gave you the idea that I suggested it? Why are you bringing this up?
I just said a little bit ago that I'm very happy with Matrox dualhead boards.
It just so happens that for good 3D performance you need a NVidia or ATI solution.
Maybe. I don't know. I suspect you don't know either because you seem to think there is no other way to get 3D performance except with on-board video.
You suggested that the only 3D applications I could run with Free drivers are some "GLQuake" and "three dee pinball" - and franky, I have no idea what you're talking about. Not because I don't know what these things are, but because I don't know why you think this. I don't know what you're talking about because you're not making an effort to be understood.
You're so convinced that you are right and I am wrong that you do the very thing that you accuse me of- arrogence and smartassness. You want to turn this into a personality issue, and presume of some kind of allusion I might have. That's generally what people do when they don't know what they're talking about.
contrary to your belief, I'm not sacrificing any freedom by not being able to modify the source of a driver
Yes you are. Read your sentence again. You are not being able to modify the source of the driver. Right now you're free to replace the oil in your car. Whether you know how to do it has nothing at all to do with whether you're free to do it or not.
Even if you don't care about your own freedom, I hope you understand that a very large number of people do care about your freedom.
So you say that Linus might not be one of them, but without Free Software, and without the FSF, Linux wouldn't even exist. Neither would FreeBSD- after all, what use would it be without a Free compiler in which to build it?
You wouldn't have Mozilla/Firefox either- because where would be their motivation to use a community that didn't exist? Microsoft wouldn't produce MSIE 7, because after all, once they had their monopoly, why would they continue working? They even promised that MSIE 6 would be the last release of Internet Explorer before Mozilla/Firefox took off!
UNIX might not have survived- after all, with SUN being the only major non-free commercial UNIX player these days, how is it they would survive when everyone else was running Windows?
And with no more UNIX, where would Apache or SSH be except without UNIX to run on, and UNIX to connect to?
Of course, Sony's PS3, and TiVO wouldn't stand a chance- No Free Software means no operating system or development kits for these guys.
Your fancy Mac OS X wouldn't be here if it weren't for Free Software- the NeXT project used a lot of Free Software for their build toolchain and their kernel!
No NeXT and no UNIX means no World Wide Web! That's right, no slashdot- we wouldn't even be having this conversation without Free Software.
So maybe you are a true pragmatist- maybe you don't have to care about your Freedoms. I guarantee however, that you've benefitted from them. You may be willing to let others erode your Freedoms, but if you are such a true pragmatist: Don't tell other people to give up their freedoms- don't try and convert them to your own foolish ways.
You might be surprised just what kind of world we'd have.
Seriously. Quick experiment. Load the "radeon" driver, and run Tuxracer. It should look like a slide show. Load the "fglrx" driver, and the same game. It should run fine.
I think it's pretty well understood that ATI isn't making the job of open-source drivers any easier. Some open-source drivers are better than others...
Anyway, going from running proprietary video card drivers to running Windows is a fairly big leap.
I'm not so sure anymore. I used to think that people were grateful for Free software, but just look elsewhere on this thread- People are honestly saying that Free Software developers should be grateful to ATI and NVidia! This is absolutely amazing to me- because that's exactly what Microsoft is saying: Be grateful to us that we even still let you use your computer.
It's one thing to not care about your own Freedom and go through the contortions to use ATI and NVidia's drivers, but it's an entirely different thing to say that other people should give up their Freedom so that they they don't have to go through these contortions.
Now, it can be noted that nVidia/ATI's drivers are distributed as an installer that compiles the module on a persons system, but also as a precompiled module that one can get through apt-get or other means. Either way ones kernel is never altered. The nVidia/ATI drivers never alter the kernel source or kernel image on disk. They are strictly modules.
No. They're not modules. They are source codes and build-routines in the form of a patch. The patch is derived from the kernel, but it isn't GPL because being derived doesn't cause a problem until it's linked.
Linking is required to turn the patch/build-routines into a kernel module, and that occurs only on your machine.
If I couldn't get them, I couldn't use them. That's why I said that IN THIS CASE distribution =~ usage.
Actually, you said distribution = usage. In any event, NO IT DOESN'T.
I have several legitimate copies of Windows NT. They are DEMO disks. When I install them, I patch them using a patch that was distributed to me by someone else. This patch removes the demo-like restrictions.
The DEMO disks are legal. The patch is legal. All aspects of distribution are legal. I can even use it legally. But I cannot take my patched system and distribute it.
Why should Kororaa stop distributing their LiveDVD? I don't think they have to, I don't feel they are at fault. And I have no idea what you're talking about anymore.
Because they're linking/creating the modules and putting the modules on the LiveCD. They are not putting the patch/build-routines on the LiveCD. They are violating the GPL.
That's cool, I'm sure you're having lots of fun playing stuff like GLQuake on your integrated video. Most people would rather play the modern games they paid hard-earned money for. Don't get me wrong, I'd love to have open source drivers that give decent hardware accelerated 3D performance, but I'd also really like to play games that take more than your standard 3D pinball game.
Forgive us for being practical.
Then be practical: Run windows. You can pay microsoft for the privilege of using your hardware just like you're paying NVidia for the privilege of using your hardware.
You may think that it's not costing you anything, but it's really costing you your freedom.
That's why Linux doesn't "get down on their knees" to NVidia or ATI- because freedom is much more valuable. In fact, without that freedom, Linux wouldn't even exist.
For what it's worth, I have no idea what this "three-dee pinball" or "glquake" you're talking about is. I don't know what kind of games you're talking about that don't work with Free GL drivers.
The way *I* learned it was that the binary blob was the Windows core driver, and the glue layer provided entry points that the Linux kernel expected.
I'd buy that if the GPL-code provided a Win32-style API to the binary blob, or even emulated FreeBSD.
Since NVidia can legally distribute their code without falling under the GPL, I would be quite confident in my own ability to do so as well
You can distribute your own code. NVidia says you can redistribute their code. This works because the NVidia "driver" is really a "patch".
Even if I distribute it as a binary kernel module in/lib/modules somewhere, it's still not a derived work of Linux. If it is, then NVidia's got a big, big problem.
Why? I can patch Windows NT "demo disks" all day, and I can distribute my patches and grant people redistribution- DESPITE the fact that it's derived from Windows.
Deriving only has an effect once it's turned into a kernel module (by linking it), and THAT is why you cannot redistribute linked kernel modules that are bound to some binary-only blob.
Distributing a Linux kernel with code, the Linux interface module with code, and the NVidia binary module without code fully satisfies the GPL, because the NVidia binary blob is not derivative of any GPL code. Period.
Says you. Not so says NVidia, and Not so says Linus. That's why nobody distributes a Linux kernel image with the NVidia binary module built for it.
Well, except Kororaa. But that's exactly why they're in trouble.
NVidia is supplying a patch, and they say you can redistribute it (under certain terms- much more restrictive than the GPL, btw). You can patch your Linux system with it. You cannot distribute the patched system. The only thing that can make it possible to distribute the patched system is NVidia must make their driver open-source (well, GPL compatible). Linus can't do it, I can't do it, you can't do it, only NVidia can do it.
NVidia supplies a handy tool for patching easily- it patches pretty easily, but you have to re-run it every time you reboot (i.e. when you replace your kernel with another version).
But it's still just a patch. The ONLY REASON they have a GPL'd glue component is to defeat the Linux kernel versioning. There is no other reason- legal or otherwise.
By your argument, the LinuxAnt people, who get wireless drivers working by the same method (an opensource wrapper on Windows drivers), are also illegal because they're not distributing the source code for the Windows drivers. That's a rather unique and novel theory, as far as I can see.
Why is that my argument?
GPL only applies to redistribution. Are these LinuxAnt people redistributing GPL'd Linux code? Or are they just distributing their patches (the opensource wrapper)?
Where is this violation of GPL that you are talking about?
There is no script. That's the problem. They're linking the binary modules themselves and distributing those.
If someone adds this software that I provide (the binary, the extra interface file and the script) to their favorite OS distribution of the month, which happens to be GPLed, are they in violation of the license, when they distribute a disk with the OS and my code?
That depends. Everyone likes fictional examples because it's easier to point out gray area, but it almost always misses some critical component of what's happening out here, in reality.
In this case, they're not adding the software that you provide (and presumably allow redistribution of), but instead, running the script, and taking the output and redistributing THAT. In this case, yes, they have committed a copyright violation, and since in this case, that thing is GPL, they are also committing a GPL violation.
In other cases? Well, you didn't explicitly say you allow redistribution, and while I assumed it for the above argument, now I won't, and yes, the person redistributing your program is in copyright violation- yours.
I find it's easier to think of binary-only drivers as patches, because it's very well understood (and tested) how patches work. Your patches can be distributed separately- and easily- just as NVidia and ATI distribute their own patches.
Anyone can apply your patches, and if you so allow- those patches can even be redistributed.
But: nobody can take the thing you patch, apply the patch, and redistribute the result. You don't have the right to do it - the copyright owner of the patch, and Linus doesn't have the right to do it- the copyright owner of Linux (well, one of many).
Usage and distribution are essentially the same in this case. nVidia and ATI make the drivers. IF they are derived works, they can't legally distribute them without the source code which means that I can't legally GET it and thereby can't possible even consider using it.
You have no idea what you're talking about. Usage is what you do on your own computer. Distribution is if you send those files you build and make to someone else. Redistribution is if THEY send those files you build and make to someone else altogether.
Since ATI and nVidia have been distributing their drivers for quite some time and there hasn't been any suit or anything, I can only assume that those who are lawyers have decided it's not a derived work.
ATI and NVidia aren't redistributing. GPL applies to redistribution. If the ATI and nVidia drivers ARE derived works, why are we going after the guy making the LiveCD and not the companies that are violating the GPL by distributing derived works.
Let's put it this way. I can take a Windows NT "demo disk" that's "valid for 30 days" or whatever, and patch it so that it's completely full featured. I can even distribute my patches and tell people where these demo disks are and how to turn them into valid Windows NT copies.
That's completely, and 100% legal, and nothing Microsoft says in their EULA can change that.
However, what I cannot do is distribute my patched Windows NT. I can distribute demo disks that I picked up, with the patches- but I cannot apply it for them.
I can make a fancy boot disk that applies my patches at install time, but I cannot distribute my patched disk.
NVidia/ATI's drivers is better thought of as a patch, and NOT as a driver or a program or a blob of any other variety.
Patching doesn't and cannot invalidate the GPL because it doesn't have anything to do with the GPL- the GPL only affects redistribution and only GRANTS RIGHTS. At no point does the GPL take ANY rights away.
I think what you're missing is that that Nvidia and ATI have a nice little duopoly going. They both can afford to reverse-engineer the "Make Doom3 Fast" function, but they aren't going to hand that information over to a competitor with a lower cost structure. Drivers are THE thing that keeps other players out of this market.
I only wish ATI or NVidia would admit to that, because it's the very definition of antitrust:)
Slashdot Mirror
Also, I want this functionality on all operating systems. Unless I explicitly request deletion, no file should ever be unlinked, deleted, or whatever you call it when I delete it, whether through the command line or the GUI.
The problem with this is, is that it causes a significant reduction in performance.
Ideally, the operating system chose the best possible spot for that file when it got written. Once that file is deleted, that spot will once again be the fastest best possible spot- for at least something. If the operating system skips that spot for a new file, then this new file isn't going to be accessed quite as quickly.
Truly automatic tiered storage solves this problem by splitting the directory services from the storage system- that is, the file's _name_ is no longer tied to the volume that the file happens to live on (and no, this isn't the same thing as symlinks or shortcuts). This allows the decision as to what the best spot for a file is to be deferred until later- and even spanned across multiple volumes!
Unfortunately, such a beast is very difficult- if we make a reduction in our requirements- say that performance isn't very important- or perhaps that we can stop using our computer for a few hours each evening, then it's probably possible. What we need is a new kind of file system that supports either atomic moves between disks, or a filesystem that splits the names from the storage.
A few research projects have been focused on these kinds of changes- but they all tend to break UNIX semantics (Amoeba immediately springs to mind)- and those UNIX semantics are, in-fact, the most widely used and recognized semantics for filesystems anywhere (Even Windows uses them!)-- people who develop a filesystem incapable of supporting them, really need to have a real good reason for breaking everyone's hard work.
While they often do, it hasn't yet been seen as good enough for general purpose stuff.
SFU: Let devs, systems admins port applications
.Net Runtime as EMCA standard
SFU: For talking to *nix clients using nfs, ldap and interact with AD
Hah. You've never actually used SFU, have you? NFS support doesn't work unless you authenticate against AD- that means making your AD server your NIS server. SFU's "porting kit", btw, consists of (get this) GCC, CYGWIN, and a number of other free softwares.
RTF Format: As an open spec that predates everything else in the opensource world
Ah, no. TeX predates RTF by about 10 years.
Check again. The ECMA submitted runtime lacks the entire WinForms interface. That is, if all you want to do is make text-based hello world programs, that's fine, but it's a far cry from usefulness.
Open XML as EMCA standard
It's ECMA. Again, OpenXML is worthless. It's got almost zero marketshare. I wouldn't want to pick up that tar baby either.
Wix install set: For open software to create installers for windows
But only ON windows. It relies on API that aren't public to do so. This negates cross-compilers, and makes farm-work difficult.
IBuySpy Portal(Dotnetnuke is based on)
Again, zero marketshare.
It is easy to blame when you look at only one point of view
I agree.
Sure MS should be releasing docs for smb, cifs, AD, rdp(dont know if they actually control it) etc.
They indeed should be! See below:
should they for free (open) that is debatable....
No it's not. They were so ordered by a US Federal Court. They are convicted criminals, so it's not debatable at all. They continue to break the law, and they continue to hurt Americans and the world at large by continuing to have turd-covered shills pretend that Microsoft is some kind of company. It's not. It's an illegal monopoly. Move on.
Where as, if there is no backwards compatibility, you are more likely to make games for the new platform than the old.
Or not at all?
Given the cost of Sony's PS3 developer package, this isn't as insignificant a problem as you might think.
I, like many people, like capturing with ethereal rather than tcpdump because it's more convenient. To do this it asks for the root password.
/usr/bin/ethereal on some systems, so that when you run /usr/bin/ethereal (a program called ethereal, but not actually ethereal) that is a symlink to userhelper which is a setuid tool that becomes root, then asks for your password, and then runs /usr/sbin/ethereal.
I, like many people, use packet sniffers to analyze network traffic besides the traffic that my workstation might be connected to (which happens to be a switch).
So what? Ethereal has lots of non-privileged user uses.
Microsoft Outlook does not need to be run as an administrator. Palm desktop does not need to be run as an administrator. What is your point here?
In less so a way that Ethereal needs to be run as root. Palm desktop will not run as non-admin without registry hacks and permission effects. Outlook won't run as a regular user unless it has been run on the machine as the administrator.
but like I originally said an administrator can cache credentials - this means you can give a user an icon on a desktop that will run an app as any user including the administrator. Sure it's a pain, but it's due to poorly written software. If I deal with software like this I'll usually look for a suitable alternative first.
So what? My point is that "cached credentials" are not like sudo or su at all. They work on a different principal, and if you allow "one icon" to run as administrator, another application can steal those credentials.
Forgive my being rusty but I don't do a lot of suexec
That's obvious.
Some applications do - ethereal for example will ask for the root password if you are not root when you run it, giving you the option to run it as unpriviliged, but it does ask the user.
No it doesn't. That's a fundimental difference between UNIX versus Windows security models. Ethereal doesn't ask for the root password- userhelper or sudo asks for it- these are setuid programs that masquarade as
Under Windows, ethereal can actually ask you for your password, and having it, it can obtain an authentication token from the system.
This is a very significant difference between the two systems- you say it may not look different, but the security implications are enormous. One way is very easy for administrators and developers to make and be safe, the other is a literal bomb that goes off regularly, and that administrators are at the whim of developers that think it's the job of the administrators.
The original code was proprietary - the folks at Berkeley kinda sorta stole it - didn't they?
No they didn't.
Even if the original Bell lab guys wanted to share, was it theirs to share?
Actually, AT&T shared it because legal obligations prevented them from entering the Operating Systems as a market.
But back in the Bell Labs days, nobody cared. It just wasn't thought about- after all, why would someone buy hardware that they can't do anything with?
That's kind of my question these days- why do people buy hardware that they can't do anything with?
I'm sure I missed some things, but other posters will point them out.
Actually yeah, the BEST way to find out how to get an application to install/run with reduced privileges is to instead search for how to install it in a Domain. The procedures are the same, but in the example of Palm Desktop- one will tell you how to do it and the other will not.
but there are many *NIX apps that expect to have root - ethereal for example.
... they are usually system admin related ...
Uh, no? I generate the dumps with tcpdump and load them into a ethereal running as non-root all the time.
Which I think is the critical difference. Microsoft Outlook has _zero_ to do with system administration. Palm Desktop has _zero_ to do with system administration.
In Windows you can use the runas command similar to su to give elevated privs to individual apps.
Well, you might say that runas is similar to su or sudo, the difference is that people don't want to, nor should they be using RunAs to open Microsoft Word. It's just stupid.
You can also use a switch to cache credentials (like chown +x root)
Err, did you mean chmod u+s?
On Windows, using cached credentials means the user session can gain Administrator privileges, but on UNIX it means that program can be run with elevated privileges.
On one hand, as long as a program doesn't ask for Administrator privileges, it'll be run in a reduced environment, but on UNIX the programs cannot "ask" to have their privileges elevated.
You seem to have confused copyrighting with software licensing
You seem to believe that software licensing exists in the absense of a signed license.
It doesn't.
You're assuming, of course, that no commercial venture would ever challenge MS for web browser domination
No I'm not. I'm assuming that they would fail. I'm assuming this because several companies did challenge MS and lost.
You recall, of course, that Unix was not open source originally, so one could claim that without AT&T there would be no open source.
UNIX most certainly was Open Source originally. See John Lion's commentary on the Sixth Edition UNIX Operating system for more details.
If it weren't for Microsoft, it would've stayed that way.
Nevertheless, I'm not talking about Open Source, I'm talking about Free Software. If UNIX was originally Free Software, it always would be Free Software. The distinction is critical.
You're full of shit because I can still use older MS products.
Software isn't like an automobile or a gallon of milk it doesn't rot. When it comes out of the factory, it doesn't begin to break down and rust, but actually any bugs that are in it are in-fact defects in the software itself.
Security bugs, while a special class, are bugs and are in-fact defects.
So if you could, back in 1998, run Windows 98 directly connected to the Internet comfortably, read your email, and work on your documents, why can't you do it now?
Of course Windows 98 had bugs in it, and of course it still does, but in order to have the productivity that they had in 1998, they can't use the software and the hardware that they had in 1998.
That's the behavior that people are referring to by "forced upgrade".
I don't give one rat's ass if I can see the source code or not
But you do care if I can see the source code. Because I am a developer, and a Free software developer at that.
If Free Software wasn't available, MSIE would still be sitting at 6.0- no PNG and no SVGs- after all, no Free software would mean no Firefox. No competition means no updates- remember before Firefox Microsoft said that MSIE 6 would be the last release!
But we don't have to stop there: If Free software weren't available, you wouldn't have your World-Wide-Web because there wouldn't be a GCC and without a GCC there wouldn't be a NeXT and without a NeXT there wouldn't be a WorldWideWeb.
Source code matters even to non-developers- EVEN when they don't realize that it matters.
Just a week ago I tried to install CVS or subversion to make a Latex document repository, after reading the fucking manual and downloading this Explorer shell extension programs (dont remember their name) I got fed up and gave up, I supposedly made a repository directory and tried to create a new tree or whatever it is called but the darn shit just could not work.
That's very brave- All those millions of people out there who do manage to get CVS to work must simply be smarter than you.
Generally, it's a good idea to ask questions when you don't know the answers.
My employer managed to install TortoiseCVS with no difficulty, and uses it daily to manage his MS-Word documents. If you like, I can give you his email address and you can ask him what he did.
And it is not really that slow and unreliable that for a lot of companies who preffer to run a setup wizzard instead of trying to figure out how to setup CVS or SV.
Again, see, you really should ask questions when you don't know what you're talking about.
CVS can and was setup via "a wizard". Module creation proved to be no difficulty- and I don't think my employer opened the manual at all.
I managed to install CVS via Debian's DPKG system and get a centralized repository setup- all from Debian's wonderful wizard-based system.
So really, while the reason you didn't install CVS or SV is simply that you're not smart enough to, that doesn't need to be a stumbling block. There are lots of people around that are smart enough to set up CVS or SV, and some of them might be quite willing to answer your questions, and help guide you through it.
Of course that's a legitimate use.
I bet the EULA and/or license for the game forbids you from using such software.
So what? You read this copyrighted post, and to make use of it (understand it), you must agree to these terms: I forbid you from eating grapes!
Does that mean it's illegal to eat grapes? You did read my post, and it's copyrighted, and it has affected you in exactly the same way as any other copyrighted material you otherwise legally receive.
So why are you still legally allowed to eat grapes? Because I can't take away your rights "just by saying so", and neither can software publishers.
but that doesn't mean I want my desktop in Linux to be hobbled because someone thinks it's a cardinal sin to taint the kernel.
The reason for "tainting" messages is so that you understand that you're giving up support. Nobody can help you with a tainted kernel except the people who tainted it (ATI or NVidia in this case).
Why? Because kernel structures have been modified and nobody knows how or why. They don't have source code, and so therefore it's very very difficult to find out where ATI or NVidia screwed it up.
And it's common sense for Pete's sake that onboard video is going to suck hard compared to a full-blown and fully featured dedicated video card
So what? I'm not suggesting on-board video. What gave you the idea that I suggested it? Why are you bringing this up?
I just said a little bit ago that I'm very happy with Matrox dualhead boards.
It just so happens that for good 3D performance you need a NVidia or ATI solution.
Maybe. I don't know. I suspect you don't know either because you seem to think there is no other way to get 3D performance except with on-board video.
You suggested that the only 3D applications I could run with Free drivers are some "GLQuake" and "three dee pinball" - and franky, I have no idea what you're talking about. Not because I don't know what these things are, but because I don't know why you think this. I don't know what you're talking about because you're not making an effort to be understood.
You're so convinced that you are right and I am wrong that you do the very thing that you accuse me of- arrogence and smartassness. You want to turn this into a personality issue, and presume of some kind of allusion I might have. That's generally what people do when they don't know what they're talking about.
contrary to your belief, I'm not sacrificing any freedom by not being able to modify the source of a driver
Yes you are. Read your sentence again. You are not being able to modify the source of the driver. Right now you're free to replace the oil in your car. Whether you know how to do it has nothing at all to do with whether you're free to do it or not.
Even if you don't care about your own freedom, I hope you understand that a very large number of people do care about your freedom.
So you say that Linus might not be one of them, but without Free Software, and without the FSF, Linux wouldn't even exist. Neither would FreeBSD- after all, what use would it be without a Free compiler in which to build it?
You wouldn't have Mozilla/Firefox either- because where would be their motivation to use a community that didn't exist? Microsoft wouldn't produce MSIE 7, because after all, once they had their monopoly, why would they continue working? They even promised that MSIE 6 would be the last release of Internet Explorer before Mozilla/Firefox took off!
UNIX might not have survived- after all, with SUN being the only major non-free commercial UNIX player these days, how is it they would survive when everyone else was running Windows?
And with no more UNIX, where would Apache or SSH be except without UNIX to run on, and UNIX to connect to?
Of course, Sony's PS3, and TiVO wouldn't stand a chance- No Free Software means no operating system or development kits for these guys.
Your fancy Mac OS X wouldn't be here if it weren't for Free Software- the NeXT project used a lot of Free Software for their build toolchain and their kernel!
No NeXT and no UNIX means no World Wide Web! That's right, no slashdot- we wouldn't even be having this conversation without Free Software.
So maybe you are a true pragmatist- maybe you don't have to care about your Freedoms. I guarantee however, that you've benefitted from them. You may be willing to let others erode your Freedoms, but if you are such a true pragmatist: Don't tell other people to give up their freedoms- don't try and convert them to your own foolish ways.
You might be surprised just what kind of world we'd have.
Seriously. Quick experiment. Load the "radeon" driver, and run Tuxracer. It should look like a slide show. Load the "fglrx" driver, and the same game. It should run fine.
I think it's pretty well understood that ATI isn't making the job of open-source drivers any easier. Some open-source drivers are better than others...
Anyway, going from running proprietary video card drivers to running Windows is a fairly big leap.
I'm not so sure anymore. I used to think that people were grateful for Free software, but just look elsewhere on this thread- People are honestly saying that Free Software developers should be grateful to ATI and NVidia! This is absolutely amazing to me- because that's exactly what Microsoft is saying: Be grateful to us that we even still let you use your computer.
It's one thing to not care about your own Freedom and go through the contortions to use ATI and NVidia's drivers, but it's an entirely different thing to say that other people should give up their Freedom so that they they don't have to go through these contortions.
Now, it can be noted that nVidia/ATI's drivers are distributed as an installer that compiles the module on a persons system, but also as a precompiled module that one can get through apt-get or other means. Either way ones kernel is never altered. The nVidia/ATI drivers never alter the kernel source or kernel image on disk. They are strictly modules.
No. They're not modules. They are source codes and build-routines in the form of a patch. The patch is derived from the kernel, but it isn't GPL because being derived doesn't cause a problem until it's linked.
Linking is required to turn the patch/build-routines into a kernel module, and that occurs only on your machine.
If I couldn't get them, I couldn't use them. That's why I said that IN THIS CASE distribution =~ usage.
Actually, you said distribution = usage. In any event, NO IT DOESN'T.
I have several legitimate copies of Windows NT. They are DEMO disks. When I install them, I patch them using a patch that was distributed to me by someone else. This patch removes the demo-like restrictions.
The DEMO disks are legal. The patch is legal. All aspects of distribution are legal. I can even use it legally. But I cannot take my patched system and distribute it.
Why should Kororaa stop distributing their LiveDVD? I don't think they have to, I don't feel they are at fault. And I have no idea what you're talking about anymore.
Because they're linking/creating the modules and putting the modules on the LiveCD. They are not putting the patch/build-routines on the LiveCD. They are violating the GPL.
That's cool, I'm sure you're having lots of fun playing stuff like GLQuake on your integrated video. Most people would rather play the modern games they paid hard-earned money for. Don't get me wrong, I'd love to have open source drivers that give decent hardware accelerated 3D performance, but I'd also really like to play games that take more than your standard 3D pinball game.
Forgive us for being practical.
Then be practical: Run windows. You can pay microsoft for the privilege of using your hardware just like you're paying NVidia for the privilege of using your hardware.
You may think that it's not costing you anything, but it's really costing you your freedom.
That's why Linux doesn't "get down on their knees" to NVidia or ATI- because freedom is much more valuable. In fact, without that freedom, Linux wouldn't even exist.
For what it's worth, I have no idea what this "three-dee pinball" or "glquake" you're talking about is. I don't know what kind of games you're talking about that don't work with Free GL drivers.
Have you checked? Do you know for certain?
The way *I* learned it was that the binary blob was the Windows core driver, and the glue layer provided entry points that the Linux kernel expected.
/lib/modules somewhere, it's still not a derived work of Linux. If it is, then NVidia's got a big, big problem.
I'd buy that if the GPL-code provided a Win32-style API to the binary blob, or even emulated FreeBSD.
Since NVidia can legally distribute their code without falling under the GPL, I would be quite confident in my own ability to do so as well
You can distribute your own code. NVidia says you can redistribute their code. This works because the NVidia "driver" is really a "patch".
Even if I distribute it as a binary kernel module in
Why? I can patch Windows NT "demo disks" all day, and I can distribute my patches and grant people redistribution- DESPITE the fact that it's derived from Windows.
Deriving only has an effect once it's turned into a kernel module (by linking it), and THAT is why you cannot redistribute linked kernel modules that are bound to some binary-only blob.
Any suggestions for a decent PCI video card with GPL driver?
I'm very happy with Matrox G400/G450 boards. I bought a bunch cheap (7$USD per board) and they're all dual-head boards so they're great at my office.
Others on this thread have pointed out Intel boards come with open source drivers.
My laptop uses a neomagic board, and I'm happy with it as well.
My home computer had a SIS-based board, and my coworker had a S3-based system.
ALL of these work with Debian Sarge, just fine.
Distributing a Linux kernel with code, the Linux interface module with code, and the NVidia binary module without code fully satisfies the GPL, because the NVidia binary blob is not derivative of any GPL code. Period.
Says you. Not so says NVidia, and Not so says Linus. That's why nobody distributes a Linux kernel image with the NVidia binary module built for it.
Well, except Kororaa. But that's exactly why they're in trouble.
NVidia is supplying a patch, and they say you can redistribute it (under certain terms- much more restrictive than the GPL, btw). You can patch your Linux system with it. You cannot distribute the patched system. The only thing that can make it possible to distribute the patched system is NVidia must make their driver open-source (well, GPL compatible). Linus can't do it, I can't do it, you can't do it, only NVidia can do it.
NVidia supplies a handy tool for patching easily- it patches pretty easily, but you have to re-run it every time you reboot (i.e. when you replace your kernel with another version).
But it's still just a patch. The ONLY REASON they have a GPL'd glue component is to defeat the Linux kernel versioning. There is no other reason- legal or otherwise.
By your argument, the LinuxAnt people, who get wireless drivers working by the same method (an opensource wrapper on Windows drivers), are also illegal because they're not distributing the source code for the Windows drivers. That's a rather unique and novel theory, as far as I can see.
Why is that my argument?
GPL only applies to redistribution. Are these LinuxAnt people redistributing GPL'd Linux code? Or are they just distributing their patches (the opensource wrapper)?
Where is this violation of GPL that you are talking about?
There is no script. That's the problem. They're linking the binary modules themselves and distributing those.
If someone adds this software that I provide (the binary, the extra interface file and the script) to their favorite OS distribution of the month, which happens to be GPLed, are they in violation of the license, when they distribute a disk with the OS and my code?
That depends. Everyone likes fictional examples because it's easier to point out gray area, but it almost always misses some critical component of what's happening out here, in reality.
In this case, they're not adding the software that you provide (and presumably allow redistribution of), but instead, running the script, and taking the output and redistributing THAT. In this case, yes, they have committed a copyright violation, and since in this case, that thing is GPL, they are also committing a GPL violation.
In other cases? Well, you didn't explicitly say you allow redistribution, and while I assumed it for the above argument, now I won't, and yes, the person redistributing your program is in copyright violation- yours.
I find it's easier to think of binary-only drivers as patches, because it's very well understood (and tested) how patches work. Your patches can be distributed separately- and easily- just as NVidia and ATI distribute their own patches.
Anyone can apply your patches, and if you so allow- those patches can even be redistributed.
But: nobody can take the thing you patch, apply the patch, and redistribute the result. You don't have the right to do it - the copyright owner of the patch, and Linus doesn't have the right to do it- the copyright owner of Linux (well, one of many).
We should note that Intel's "new hardware" is inferior to what you consider ATI's "old hardware"...
Except in the way that really matters: Intel provides open source drivers and documentation.
So if OpenOffice were GPL, then all Word documents are subject to the GPL?
I think you are seriously reaching here.
No, your analogy just sucks.
Word documents could not possibly be retroactively derive from OpenOffice.
Furthermore, Unless the document is about the program, I don't see how its possible to derive a document from a program.
Usage and distribution are essentially the same in this case. nVidia and ATI make the drivers. IF they are derived works, they can't legally distribute them without the source code which means that I can't legally GET it and thereby can't possible even consider using it.
You have no idea what you're talking about. Usage is what you do on your own computer. Distribution is if you send those files you build and make to someone else. Redistribution is if THEY send those files you build and make to someone else altogether.
Since ATI and nVidia have been distributing their drivers for quite some time and there hasn't been any suit or anything, I can only assume that those who are lawyers have decided it's not a derived work.
ATI and NVidia aren't redistributing. GPL applies to redistribution.
If the ATI and nVidia drivers ARE derived works, why are we going after the guy making the LiveCD and not the companies that are violating the GPL by distributing derived works.
Let's put it this way. I can take a Windows NT "demo disk" that's "valid for 30 days" or whatever, and patch it so that it's completely full featured. I can even distribute my patches and tell people where these demo disks are and how to turn them into valid Windows NT copies.
That's completely, and 100% legal, and nothing Microsoft says in their EULA can change that.
However, what I cannot do is distribute my patched Windows NT. I can distribute demo disks that I picked up, with the patches- but I cannot apply it for them.
I can make a fancy boot disk that applies my patches at install time, but I cannot distribute my patched disk.
NVidia/ATI's drivers is better thought of as a patch, and NOT as a driver or a program or a blob of any other variety.
Patching doesn't and cannot invalidate the GPL because it doesn't have anything to do with the GPL- the GPL only affects redistribution and only GRANTS RIGHTS. At no point does the GPL take ANY rights away.
No, but if you use ubuntu you can get the compiled GPL module in the restricted modules package and the binary part in the nvidia driver package
Look inside those packages. They actually download and link at install time. No distribution takes place.
I think what you're missing is that that Nvidia and ATI have a nice little duopoly going. They both can afford to reverse-engineer the "Make Doom3 Fast" function, but they aren't going to hand that information over to a competitor with a lower cost structure. Drivers are THE thing that keeps other players out of this market.
:)
I only wish ATI or NVidia would admit to that, because it's the very definition of antitrust