![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
lprovenJust a passing thought, prompted by something ![[livejournal.com profile]](https://www.dreamwidth.org/img/external/lj-userinfo.gif)
vicarage said in the L_O_L LJ.
Even though the Unix wars are over and there's been much consolidation, there are still lots of Unices about. Linux is big, NetBSD/OpenBSD/FreeBSD seem to be doing OK and DragonflyBSD is going its own way, Solaris and AIX and HP-UX persist in the commercial space and Solaris is moving towards being a commercially-sponsored FOSS OS. And there's OS X with 10s of millions of users, which is Mach+bits of BSD + bits of NeXTstep & OpenStep + a proprietary GUI).
Even within Linux, there's much variability: some of the main axes, as I see it, are:
RPM distros versus Debian distros (e.g. Red Hat/SUSE/Mandrake v. Debian+derivs & Ubuntu+derivs)
Commercial distros versus free ones (e.g. Redhat, SUSE, Mandriva, Xandros & Linspire &c. v. Debian, Ubuntu, Slackware &c)
SysV init versus BSD init (everyone versus Slackware+derivs plus 1 or 2 others)
Distros which have done a deal with MS versus ones that haven't (Novell SUSE, Linspire/Freespire & Xandros versus everyone else)
... and so on.
So there's still an awful lot of variation out there. It's far from FOSS versus commercial. The lines are very blurred.
But Linux wouldn't exist if it wasn't for GNU. Indeed RMS still wants people to call it GNU/Linux, with some justice. GNU was meant to have its own kernel: initially TRIX from MIT, then they considered BSD 4.4-Lite, then they spent 15y working on a Mach-based microkernel, then since 2005 or so, much effort has gone into work based around the newer L4 m/k, while some people are looking at Coyotos (which - along with CapROS - is based on EROS, which derived from KeyKOS).
Summary: it's complicated, it's really hard and they're not getting far.
But what if, back in 1989 or so, GNU had gone with BSD 4.3 Net/1? They only needed the kernel, really - they were working on their own userland. As the guy who wrote HURD in the first place said, "My first choice was to take the BSD 4.4-Lite release and make a kernel. I knew the code, I knew how to do it. It is now perfectly obvious to me that this would have succeeded splendidly and the world would be a very different place today."
Even way back in '91, Linus famously said that his pet project "won't be big and professional like gnu". If he'd had a usable GNU system back then, I think he's said he'd never have bothered.
Today, we probably wouldn't have the whole confusing set of various forms of unfinished GNU OS plus dozens of Linuxes plus half a dozen or so BSDs. If GNU had had a working kernel, it probably would never have spent all the effort on microkernels. If the BSD kernel had been the basis of a complete, usable, Free system using GNU, they might never have bothered with all the faffing about to produce a complete AT&T-code-free release.
It would probably resulted in a working free Unix a lot earlier. Depending on how you define finished and usable, Linux took 'til around '94 (kernel 1.2) to '96 (kernel 2.0).
And if the GNU free Unix had been usable by the early 90s, I reckon commercial Unix would be a lot worse off than it is today.
We might have just GNU and one or two commercial offerings by now - probably AIX and Solaris, I reckon. NeXTstep would have based on the GNU OS, probably. The Unix world would be a much simpler place. NT would probably still have happened - it was well underway in the early '90s - but it might not have got so far.
Did we miss out on a wonderful chance? Or is the modern spread of variants a good thing, showing adaptive radiation and diversity?
Discuss.
vicarage said in the L_O_L LJ.Even though the Unix wars are over and there's been much consolidation, there are still lots of Unices about. Linux is big, NetBSD/OpenBSD/FreeBSD seem to be doing OK and DragonflyBSD is going its own way, Solaris and AIX and HP-UX persist in the commercial space and Solaris is moving towards being a commercially-sponsored FOSS OS. And there's OS X with 10s of millions of users, which is Mach+bits of BSD + bits of NeXTstep & OpenStep + a proprietary GUI).
Even within Linux, there's much variability: some of the main axes, as I see it, are:
RPM distros versus Debian distros (e.g. Red Hat/SUSE/Mandrake v. Debian+derivs & Ubuntu+derivs)
Commercial distros versus free ones (e.g. Redhat, SUSE, Mandriva, Xandros & Linspire &c. v. Debian, Ubuntu, Slackware &c)
SysV init versus BSD init (everyone versus Slackware+derivs plus 1 or 2 others)
Distros which have done a deal with MS versus ones that haven't (Novell SUSE, Linspire/Freespire & Xandros versus everyone else)
... and so on.
So there's still an awful lot of variation out there. It's far from FOSS versus commercial. The lines are very blurred.
But Linux wouldn't exist if it wasn't for GNU. Indeed RMS still wants people to call it GNU/Linux, with some justice. GNU was meant to have its own kernel: initially TRIX from MIT, then they considered BSD 4.4-Lite, then they spent 15y working on a Mach-based microkernel, then since 2005 or so, much effort has gone into work based around the newer L4 m/k, while some people are looking at Coyotos (which - along with CapROS - is based on EROS, which derived from KeyKOS).
Summary: it's complicated, it's really hard and they're not getting far.
But what if, back in 1989 or so, GNU had gone with BSD 4.3 Net/1? They only needed the kernel, really - they were working on their own userland. As the guy who wrote HURD in the first place said, "My first choice was to take the BSD 4.4-Lite release and make a kernel. I knew the code, I knew how to do it. It is now perfectly obvious to me that this would have succeeded splendidly and the world would be a very different place today."
Even way back in '91, Linus famously said that his pet project "won't be big and professional like gnu". If he'd had a usable GNU system back then, I think he's said he'd never have bothered.
Today, we probably wouldn't have the whole confusing set of various forms of unfinished GNU OS plus dozens of Linuxes plus half a dozen or so BSDs. If GNU had had a working kernel, it probably would never have spent all the effort on microkernels. If the BSD kernel had been the basis of a complete, usable, Free system using GNU, they might never have bothered with all the faffing about to produce a complete AT&T-code-free release.
It would probably resulted in a working free Unix a lot earlier. Depending on how you define finished and usable, Linux took 'til around '94 (kernel 1.2) to '96 (kernel 2.0).
And if the GNU free Unix had been usable by the early 90s, I reckon commercial Unix would be a lot worse off than it is today.
We might have just GNU and one or two commercial offerings by now - probably AIX and Solaris, I reckon. NeXTstep would have based on the GNU OS, probably. The Unix world would be a much simpler place. NT would probably still have happened - it was well underway in the early '90s - but it might not have got so far.
Did we miss out on a wonderful chance? Or is the modern spread of variants a good thing, showing adaptive radiation and diversity?
Discuss.
