On 19/09/06 20:06, M Harris wrote:
On Tuesday 19 September 2006 20:45, Stevens wrote:
Why do some in the Linux community treat ndiswrapper with such contempt? It seems to me that if something works, especially as well as ndiswrapper, what damned difference does it make if the device drivers were written for use in Windows? I rather think that the concept should be expanded to include the universe of devices that are not currently useable with Linux due to unavailable drivers.
hi Fred, The problem with ndiswrapper (or any other binary driver) is that it breaks the linux kernel philosophy, and also poses a threat to the stability of linux in general. One reason Windoze is *broken* is that way too many vendors (binary drivers) are trusted to run at ring zero (root level). This is not going to happen with Linux... and thankfully so far, the linux kernel development people are sticking to their guns on this. If you want to *taint* your kernel and take the chance that that binary driver you just loaded into your kernel is safe, robust, and can be trusted... go for it. However, the open source gpl concept says that the kernel is going to be open, stable, and untainted (unless the user decides to shoot his/her own foot)
With my previous sound card, the module (native Linux) distributed by SuSE tainted my kernel. Recently, I got a couple of these: Sep 5 20:34:55 static24-89-67-198 kernel: thinkpad: module not supported by Novell, setting U taint flag. Why the system thought this was a thinkpad is beyond me, it isn't even a laptop... so much for shooting *myself* in the foot.