Hello! I have read announcements and basic descriptions of the build service, smart and of a new concept of kernel modules, so maybe my idea is already in those new concepts. Apart from the discussion around GPL or binary, I thought it might be very nice for the user, if there was an easy way to get WLAN working. There is currently yet another discussion about binary on the kernel-mailinglist due to intel's announcement of their new chipset's driver, so it seems that it becomes more and more difficult for a normal user to get its existing hardware to work. Further there is no site on opensuse, that has detailed, complete and reliable data on which chipsets will be natively supported in 10.1 and newer. One way or another, justified or not, the user will blame SuSE, if the hardware cannot be made working easily. So I want to ask, whether it would be feasible to have a "WLAN-repository", which can be used by a user without searching the web for the chipset used by one's card or getting any hints from somebody that is more familiar with that matter. Of course one should not be forced to install the driver again after each Kernel-update or have to compile anything. The only thing needed should be a hint during installation that the WLAN-card is not supported because of the drivers not being open source and that one should start "install-wlan-driver" after the installation has finished. That tool would give access to the WLAN-repository, which is fed as packman is. Since there are only a handlful chipsets, it should be possible to provide drivers for each, including a list of wlan-cards that use them. The user could either search for a chipset, or the name of the wlan-card and simply install the driver. For those cards that use a firmware, it wold be very helpful to have a script bundled with the driver that is executed by the tool and installs and renames the files needed, as YOU currently does, or maybe by searching the manufacturer's driver-cd. Please do not make this a discussion about whether GPLed-drivers would be the solution, since that will not happen, at least not until 10.1. Further there seem to be some legal issues that force companies to have at least some part of their drivers in binary, which might make not sense, yet companies like Intel simply have to comply the law, no matter whether it makes sense or not. http://thread.gmane.org/gmane.linux.network/32620 Sven