The driver Catch 22
|
| Author | Content |
|---|---|
| rob_hughes Dec 19, 2005 5:51 AM EDT |
First, let me qualify what I'm about to say. I would greatly prefer that all hardware vendors open source their drivers. That said, I doubt it's ever going to happen. The primary reason is that many vendors license 3rd party patents, code, etc., in order to speed production times of both hardware and drivers (or, in the case of patents, to keep from getting sued), and so are contractually bound to keep that secret. This is where the conflict comes in. I've seen far too many people involved in the kernel threatening to make it as difficult as possible for vendors to support linux through closed binary-only, or binary-plus-a-glue-blob, drivers in order to attempt to force them to open the driver. This is what I think is more likely to make linux fail than any other strategy, for the simple reason that if linux becomes too difficult for vendors to support in the manner they choose, they'll simply drop support, not change their basic strategy. I understand why this group of kernel guys is doing this, as they don't want to see vendors start picking and choosing among linux distributions to support. They also believe they can do a better job of resolving bugs in the drivers, or at least a faster job, and maintaining compatibility as various changes are introduced into the kernel. It would also insure that full support for all drivers is available the day a new kernel is release. And for the most part, I agree with these statements. What these same groups utterly fails to realize is that while linux kernel development may be a meritocracy, the hardware vendors are not. They just want to make as much money as possible with as little investment as possible. They'll support linux as long as it's profitable to do so. So what we end up with is a clash of two cultures, with neither side appearing to give any ground, or offer any compromise in any way whatsoever. I just hope this doesn't end up killing linux adoption. Or if not kill all adoption, at least mortally wound linux desktop adoption. And server adoption, for that matter, as there's not much point in putting linux on a server is you can't utilize the hardware to the fullest. Or worse, we could end up with a completely fractured linux landscape, with the kernel.org kernel, a RedHat kernel, a SuSE kernel, a Debian kernel, a Nvidia kernel, an IBM kernel, a Dell kernel, etc., none of which are really compatible with the others because of what bits each chooses to support and the various hacks to the kernel to get each vendor's extensions implemented. And that would, in my opinion, be the true beginning of the end of Linux adoption. I have no idea what's going to happen with all this. But with all the noise and arguing over this subject, I'm now a lot more concerned about these developments that I ever was about the SCO law suit. |
Posting in this forum is limited to members of the group: [ForumMods, SITEADMINS, MEMBERS.]
Becoming a member of LXer is easy and free. Join Us!