1 Sep
2012
1 Sep
'12
5:38 p.m.
On 2 September 2012 00:10, Dave McGuire <mcguire at neurotica.com> wrote:
On 09/01/2012 06:12 PM, Liam Proven wrote:
[Nod]
So candidates were TRIX, Sprite, BSD, UZI, then they tried to build
HURD and in the end it became Linux.
Well, to the very end of that sentence, I say "yes and no". The Linux
crowd used a lot of GNU userland stuff from the very beginning, with little
or no involvement from the GNU project, and certainly no *official*
involvement. When RMS realized (well, more like "admitted to
himself") that
the Hurd was essentially stillborn, they pretty much laid claim to Linux.
As far as I saw, they didn't go so far as revisionist historianism, but they
did try VERY hard to get everyone to call it "GNU/Linux".
Sounds about right. I was following with mild interest until the Linux
kernel passed version 1.0, at which point I started paying attention,
attempting to get the thing to install, etc.
That
riding-on-coattails thing is something they should've been pretty
embarrassed about, but they certainly didn't seem to be at the time. RMS
was not at all bashful about discussing Linux as "GNU/Linux" at every turn,
simply because they adopted the GNU userland utilities...as did several
other OSs.
I'd agree with that, but as RMS sees it (I reckon), it was built using
his tools to work with his userland; it would be nothing without the
GNU Project, and thus it's de facto /part/ of the GNU Project.
Windows, for example, doesn't somehow become
"GNU/Windows" when one
installs Cygwin. ;)
Ha! Good point!
The GNU OS project went on and on for a long time,
as the dream of several
people, so I think it was very tough for them to just throw in the towel and
be content with the fact that their *utilities* were what the world depended
on, and they didn't build the entire OS.
Well, in the longer run, I think Andy Tanenbaum was right. The
monolithic-kernel ideal was past its time in 1991 and it's
ridiculously so 20+ yr later.
Even now it's based on L4 and not Mach, I think the microkernel idea
has merit. There are others out there - Minix 3 springs to mind.
I suspect that eventually the Linux project will just run out of
steam, falter and stagnate, and maybe, just maybe, by then, the
distributed/loosely-coupled-array-of-servers type of OS will rise up
from its ancient slumber and take over the world.
I have also read that the design of Plan 9 makes the whole monolithic
kernel/microkernel debate moot, but I don't know enough about Plan 9
to judge & anyone who does is unable to talk down to my level to
explain it to me. ;?)
Of course this is all just useless "what the
hell would I know, I was only
THERE" type of information, but I could put it on a webpage if you like. ;)
All the happenings from that era, which wasn't really very long ago at all,
well less than a decade, and all of the above is pretty generally known.
The real mostly-forgotten story, as I remember it, is here:
http://www.h-online.com/open/features/GNU-HURD-Altered-visions-and-lost-pro…
... as sourced from here:
http://www.groklaw.net/article.php?story=20050727225542530
That's one of the biggest missed opportunities in the history of
computing right there.
It took 4 or 5 /years/ from 4.3BSD Net/1 to BSD 4.4-Lite, when the
real work could really begin. By then, Linux + GNU had stolen the
thunder.
If GNU had adopted the BSD kernel in the late '80s, then the whole
AT&T code issue would have gone away. (AIUI.) There might not /be/ a
BSD family of Unixes (which would be a shame) but the Free Unix world
would have got bootstrapped before Microsoft got in there.
For me, the greatest missed opportunities in the last 30y of computing
were (in approximate chronological order):
* Sinclair's Spectrum 128 *not* being based on the Timex TS2068 improved ULA
(Sinclair was focussing on the abortive QL & missed the chance to keep
the Spectrum on top of the 8-bit heap)
* Apple castrating DR GEM on x86, leading to the split from the 680x0 version
(GEM as a multitasking GUI on top of a multitasking CDOS386-based OS?
Yes please!)
* OS/2 1.x not being 386-specific, in which case, it coulda been a contender
(Allegedly IBM wanting to support its 286 customers, but visibly
stupid /even at the time/)
* GNU not adopting the BSD kernel in 1989-1990
(Stalling the development of a Free Unix for 4-5 vital years)
* Quarterdeck not getting Desqview/X out until after Windows 3.1
(DVX was fantastic, but too late)
* Apple rushing the Newton out before it was ready
(Abandoning Dylan, which could have been a truly transformative
language, I still feel)
* Apple not spotting that the answer to Pink, Copland etc. was A/UX on PowerPC
(But that would have killed the NeXT deal, which saved the company, so... moot)
* Be and BeOS getting shafted by MICROS~1
(Big players like Hitachi were bundling it but couldn't dual-boot
because of MS' OEM contract on MS-DOS & Windows)
Anyone got any other suggestions for the list? :?)
--
Liam Proven ? Profile: http://lproven.livejournal.com/profile
Email: lproven at cix.co.uk ? GMail/G+/Twitter/Flickr/Facebook: lproven
MSN: lproven at hotmail.com ? Skype/AIM/Yahoo/LinkedIn: liamproven
Tel: +44 20-8685-0498 ? Cell: +44 7939-087884