no it will not. the main reason people say hurd is technically superior is that it is a microkernel, while linux is monolithic. microkernels hand off as much responsibility to other software, to optimise speed and ease of customisability for future versions. Monolithic kernels (such as linux) contain everything in the kernel, a potential headache for anybody trying to add to it or tidy it up. But wait, linux sensibly decided the linux kernel would be totally modular, meaning each module can be developed more or less in total isolation from the other parts of the kernel, also, microkernels have a habit of being flaky in practice due to several reasons. Hurd has been vapourware for
twenty years, linux was in stable versions between a year and two years after it was conceived.
you might like to read this article about the history of GNU and Linux, in particular
read this link which is a transcript of the "Linux is obsolete thread on comp.os.minix, which answers most of these questions.
finally one quick word:
quote:
Creating his own OS based on Minix, an x86 based UNIX used to teach students to run UNIX on big servers and mainframes.
everything else you said is spot on, but this is not quite right. Minix was a proprietary OS that Andrew S Tanenbaum wrote for the 8088 PC (i think, it may have been the 8086 but i don't think so) as a teaching tool. He did it in his spare time and he charged money for it. Tanenbaum was famous amongst his users for hardly ever adding features that everybody was asking for, and for taking ages to upgrade when he did add new features. His arguments were always "i don't have time to add that" and "Minix would be too complicated a teaching tool if i added that". His users got very frustrated with him, and when the 80386 PC had been out for a while (the 286 was an utter disaster technologically and commercially), and it became apparent that Tanenbaum would not port his mnix system to the 386, some people got even more hot under the collar. Bruce Evans (later to work on linux) brought out a series of patches you could apply to Minix to make it work better on your 386, but tanenbaum did not endorse these and they were unsupported by him. Minix at this point is still the only unix you can run on a PC, all other unixes require you to spend another couple of thousand bucks on a sparcstation or something.
Enter Torvalds, he has upgraded from a Sinclair ZX to a 386 although he is not too happy with the intel chip architecture, he orders a copy of minix, and after a few months he writes a little terminal switcher because minix' multitasking is rubbish (the reason being that it was written for an 8088 and not a 386, which had better multitasking), so he could read his messages and work on his computer at the same time. Over a short time he adds more and more features to his multitasker and realises it could easily become a kernel on its own, since minix looks to him like it is frozen in time, he does, and he takes the unprecedented step of asking for help on comp.os.minix. He uses ftp (provided by ari lemmke of funet) to distribute his work, and accepts patches from people by email.
Both minix and linux are based on the cheap intel PCs, not mainframes. minix was designed to teach students about how unix microkernel systems work (with nice tidy subroutines et c) and linux was designed to be easy to develop in modules and most importantly to be workable now, with no difficult hurdles to overcome before it is usable.
Hurd is a microkernel. Most people who say microkernels are technically superior say it from an academic perspective with not much basis in what actually works. no offence to psyjax here however. microkernels
are superior to monolithic kernels technically, but try getting one to work! also Torvalds took several noteworthy decisions which allowed his monolithic kernel to avoid the problems previous monolithic kernels had fallen prey to.
[ May 13, 2003: Message edited by: Calum: hopelessly outnumbered ]
