You completely missed the point. I compared automakers and software developers as an example to demonstrate the lack of politics, not proof thereof.

You brought it up, so you tell us why.
But you are right of course. It has nothing to do with including software from another, different, system and everything to do with politics.

As a "side note", PTrenholme's post would be worthwhile reading.

I wasn't answering to you only.

Also, you particularly appear to imply that you closed the door to the possibility there are politics in linux or in Ford. You are naive if you think that.

Nothing is devoid of politics. It is a scale from very little to completely. Even choosing to use Linux or BSD instead of Windows is partly political in a small way to most users, and very political to a few. What I am "closing the door on," is the notion that not including Windows drivers is a political decision. That idea dismisses reality.

You must remember that device drivers are *OS specific* (so, no politics involved here).

As it has been multiple times demonstrated, they can be wrapped to work so they are not forced to be specific in the practical world.

Also, especially for the simpler devices, the CPU load is non-existent on modern processors.

e.g. for a personal problem involving USB sound (Creative's special effects, not base output), I doubt it would inflict much more than windows since the driver is already software based there largely.

So write one...

1 members found this post helpful.

Short answer: Yes, there is.


Linux desktop isn't really desperate for an attention, i think it's quite happy with itself. There is no need for tons of people, especially if they required windoze drivers in the kernel.

1 members found this post helpful.

How ironic that elitist view is when Linus Torvalds has multiple times demonstrated it's one of the things he doesn't like about the community.

He's a bit cute though when it goes to keeping balances. He often explodes with things like "WTF, just use KDE, it's more complete" but then retracts for a couple of years not talking about it or saying he uses Gnome now.

Get it well in your head. Torvalds is not an elitist and he likes the Desktop to succeed.

You may disagree, but I'd rather go with the side of smart people.

I'm sorry if i sounded elitist or even hostile, that was not my intention. I was merely trying to amplify the fact which was already stated - Linux is not Windows, Mac or anything else - and this IS good reason by itself, something i had the assumption you ignored the whole debate.

Not sure why you brought up the name of Linus Torvalds, i didn't mention him anywhere.

Well, that's actually the point of my previous post, the definition of "success". GNU/Linux is different, and only that can make it better. I hope your "tons of people" can understand it, and - above else - will accept it. If the linux desktop starts to make compromises just to attract people it might attract some people, but will it still be LINUX desktop?

As the previous poster said, what exactly has this got to do with Linus?

Sorry, but I'm bored of reading the ill-informed sermons from people, who only joined last month and in this case has made 13/18 posts in this particular thread, pontificating and sermonising about how GNU/Linux can be saved by supporting closed source windows drivers... If you actually understood the implications and consequences of what you're suggesting, this thread wouldn't exist. In my opinion you're simply trolling, for "political" answers in particular and have been frustrated in this so far.

It would actually be an easy thing to accomplish. The people who maintain the Linux kernel only need to place a telephone call to Microsoft's headquarters in Seattle and say, "Hey dudes. We want to include Windows drivers in our kernel."
Microsoft would gladly agree to share their patented and secret software, in order to help their Linux brothers and sisters and make Windows compatible with Linux. They would welcome Microsoft customers ceasing to buy their products and using the free competition. (Competition in their eyes.) I wonder why Linus has never thought of it?

Your sarcasm aside, perhaps it's because, when driver developers in the EU sued MS for access to the information needed to write Windows drivers for their products, and the EU Court forced them to comply, what they provided was almost unusable, and the developers had to sign nondisclosure agreements before they could see even the poor MS documentation. (Traditionally, MS has relied on its code as its documentation, so all those "hooks" in the OS needed to write good drivers are buried in the code, and -- once you find a "hook" -- you have to guess about the various parameters in the call to that "hook.")

A much easier resolution to this thread might be to ask why some vendors are not writing Linux drivers for their products. It's not, I think, anywhere as hard to do as it is to write a Windows driver.

Something occurred to me reading this thread: Might it be as difficult to port Windows drivers to Linux as it is to create Linux drivers? I always got the impression that ndiswrapper was due to having to push binary blobs into the wireless card and that that drivers could have been written for either OS?
I'm certainly no coder so I have little idea of the intricacies but it did occur to me that the reason any specific Windows driver isn't available in Linux is similar to the reason there is no Linux driver -- sadly there isn't enough developer time to produce it but in the case of the Linux driver it would be worth while.

Not exactly: what ndiswrapper does is to use the Wine project's emulation of the Windows OS so the Windows driver can be run "as though" it was running under Windows. That added emulation layer "eats" CPU cycles, and degrades the driver's performance.

Hardware providers often incorporate patented or "trade secret" components in their hardware that require special, undocumented, binary codes to be sent to the hardware to control its operation. That is what the driver does. But publishing the specifications of those hidden codes would (in the minds of the lawyers) reveal the secrets, and, therefore, make it no longer a protected secret. There are two solutions that a manufacturer may take to make a driver for a Linux system:

1. Decide that keeping the "secret" of how to use the hardware is silly, and publish the full specification and driver code. That gives the manufacturer a large pool of eyes to suggest improvements and flaws, and, usually, results in a better product.
2. The other approach is to write a program that can be called to add the "secret sauce" to the data going to the device, and decode the device's response. That binary "blob" can be called from a standard kernel module.

The second method lets the manufacture preserve its secrets and to provide a viable driver for a Linux system. That is, as you may know, exactly what nVidia has done to provide a Linux driver for their video card(s).

The "political" problem with the second method is that the "blob" is executed in the kernel, with, of course, "root" privileges. Thus any system incorporating such a "blob" in its kernel cannot be considered a secure system. Nor can it be considered an "open source" system. (That's why the Fedora project, for example, will not include an official nVidia video driver in its distributions or repositories.)

1 members found this post helpful.

How many Open Source drivers will be in the Wins8 kernel? In fact Open Source software?
Can we get a boot?