Messages in this thread | | | From | "Khimenko Victor" <> | Date | Sat, 19 Sep 1998 22:31:26 +0400 (MSD) | Subject | Re: Linux, UDI and SCO. |
| |
In <19980919193105.A22160@arbat.com> Erik Corry (erik@arbat.com) wrote:
EC> In article <ABDFz0suqH@khim.mccme.ru> you wrote:
>> But Terry still wrong. There are is GPL and even more important there is >> additional Linus clause about possibility of binary drivers for Linux. >> So we have three choices for UDI drivers: >> 1. GNU GPL'ed UDI drivers or similar. Linux community are happy. [Some] Unix >> vendors are unhappy. [Some] hardware manufacturers are unhappy. >> 2. GNU LGPL'ed UDI drivers or similar. Linux community are content. Unix >> vendors are content. [Some] hardware manufacturers are unhappy. >> 3. Closed-Source UDI drivers. Linux community is unhappy. [Some] Unix >> verndors are content, some are unhappy. Hardware manufacturers are >> happy.
EC> But you can release a UDI driver simultaneously under two EC> different licenses. So everyone can be happy.
Not at all. LGPL will be 100% enough for such purposes. Even preferred. Problem is only for manufacturers: they are effectively forced to give away specs for hardware without NDA's. This is EXACTLY what most of hardware manufacturers dislike to do.
EC> I think it is in the hardware manufacturers interest to EC> release their UDI driver under the GPL or the XFree86 EC> license because.
EC> 1) It costs them nothing. 2) It gets taken up in the EC> standard kernel, which reduces the work for them. EC> 3) It means they can sell to non-x86 (and later non-Merced) EC> Linux users EC> 4) Someone might use the UDI driver to make a native Linux EC> driver which is almost certain to have better EC> performance. EC> 5) It makes them look good. 6) Non-Linux Non-mainstream EC> OSs can use it too, without having to bother the EC> hardware manufacturer for a precompiled version (if EC> they use GPL then this applies to individual users of EC> the other OS, if they use XFree86 licensing the driver EC> can go into the standard distribution of ARM-OpenBSD EC> or whatever. EC> 7) People who insist on source for security reasons EC> or other reasons can use the hardware too. EC> 8) People will send them bug fixes, which they can use for EC> non-source UDI platforms EC> 9) Users who want to be able to file bug reports to Red Hat EC> etc. and the kernel list will be able to use their EC> hardware. I would imagine that the kernel developers EC> will not want to waste their time trying to debug EC> kernels that contain non-source UDI drivers.
EC> If the hardware manufacturer doesn't see this then we EC> can recommend users not to buy from them like we do now.
When drivers just could not be used with recend kernel versions (as of now) it's trivial. UDI will make this hard to do. Plus most Linux users in 21th century will be "Joe Averages" who will not be bothered by such obscure things and who will be sure that it's Linux fault if Linux constantly made kernel oopses even if they are using buggy binary-only drivers...
EC> I don't believe in the doomsday scenarios, on the contrary EC> I think this could usher in a new age, where drivers for EC> new hardware are available immediately instead of several EC> months later. It is often a frustrating experience EC> loading Linux on a brand new machine, because being EC> brand new it has new hardware, where the drivers are EC> untested or unavailable. Three months later it is easy EC> but people don't generally buy three month old machines. EC> That turnaround time could improve dramatically.
Yes. But this will effectively mean for hardware manufacturers that they are forced to publish specs for their hardware. I'm not think that manufacturers will be happy to do this :-(( Of course this is the only disadvantage for manufacturers but look like some of them REALLY hate to give away specs...
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
| |