On Thu, 2003-02-20 at 21:38, Theron William Stanford wrote: > So, Red Hat is forking? Bad, bad, bad. "If I really wanted to"? You mean to > say they're making decisions for me and giving me stuff that I wouldn't have > otherwise and probably don't even need unless "I really wanted" it? Sounds like > there's stuff going in that shouldn't, and I don't want a dirty kernel.
No, they're not forking. They're exercising the rights the GPL was created to protect. They feel that their customers will be benefited by the inclusion of certain features or fixes. As long as these modifications don't create incompatibilities, its not a big deal. As has been pointed out, they employ Alan Cox, Linus' former right hand man. He's a level headed, experienced developer and when it comes down to it I personally trust him before Linus. RH is modifying the kernel, but they keep going back to the original. Forking is creating a competing project. Like Wine and WineX, Samba and Samba NG. It isn't always bad, in moderation and with good reason it can benefit both parties. Witness the "synergy" of the *BSD forks as they fill a niche but share code with one another. > (For another example, I've been looking at the Mandrake init scripts, and they > keep mentioning something called Aurora. If they're going to add > semi-proprietary stuff, stuff that isn't "traditional" Unix, I want out.)[1] That I can understand totally. If Freedom really is important to you, you'll probably love Debian. > Now, I may have my Linux history wrong, but if patching the kernel is leading to > conflicts, it sounds like someone (or a group of someones) has lost control of > the whole project, or that it's being forked by the distros, and if there is no > control, from what trustworthy source do I get a decent (non-forked) kernel, > along with patches that should fit in so well that I don't have to worry about > conflicts? Conflicts in the sense that when two patches touch the same section of code, you have to go in and intelligently merge their functionality. Generally, though, they don't create incompatibilties noticable to applications. There is one correct way for a packet to be represented at the network level, and their is one standard API that must be conformed to, but in-between it doesn't matter if the implementation is zero-copy or copy like crazy--but one implementation is clearly faster. You can get vanilla kernels at kernel.org, but don't fear the diversity. Part of the wonder of Linux is that amid all this chaos real work gets done. If you feel you can write a better scheduler, you can. When you're done, other people can use it if they like it. If it's good enough, it'll find its way into the stock kernel. But if it doesn't, you're still able to meet your specific needs without writing an entire kernel. > Maybe I should use FreeBSD, or maybe the HURD kernel, or maybe Debian. Some > suggest Gentoo, but since I don't know its background, it is currently suspect. If you go exclusively HURD, you'll have nothing but my respect. (And pity until its more mature.) > > Presumably most other distributions do much the same thing. > > Which is why I'm eventually going to run Linux distro-less and enjoy true, total > freedom. > > However, I am still immensely enjoying vga=791! Hmmm, perhaps your message was meant tongue in cheek. Oh well, I can't help gushing anyway. -- Stuart Jansen <[EMAIL PROTECTED]> #define FALSE 0 /* This is the naked Truth */ #define TRUE 1 /* and this is the Light */ -- mailto.c ____________________ BYU Unix Users Group http://uug.byu.edu/ ___________________________________________________________________ List Info: http://phantom.byu.edu/cgi-bin/mailman/listinfo/uug-list
