I probably should have just said it's the old defunct religious war bullcrap
and left it at that.  The micro vs monolithic kernel was just the most
famous of his early flamewars:

https://en.wikipedia.org/wiki/Tanenbaum%E2%80%93Torvalds_debate

Linus has never really given a consistent answer as to why he turned his
back on the efforts going on with BSD and gone with the Minux rewrite.

One day it's the micro/max kernel.  Another day it's the USL lawsuit.  Still
another day it's something else.

Frankly IMHO Linus is just a guy out of many back then who were putting
development time into Unix-alike systems and through sheet LUCK his crap got
a critical mass of developers and users formed around it before the other
Unix-alike and descendants of actual Unixes did.  And now his net worth is
$50M and he runs around pontificating and everyone has to pay attention to
his spewing, while people like William and Lynne Jolitz who certainly did as
much work as he did back then, are mostly forgotten.

And, I think he knows it.  And I frankly think he's been sort of bewildered
why he almost randomly got selected.

Linux for all it's engineering, would have been nothing more than a
historical footnote if it wasn't for all the other developers contributing
to it and most of all, all the users using it.  And it was 99% luck that
this critical mass happened with Linux at the right time - there were other
efforts in Unix going on back then, some of them arguably more mature.
386BSD for example waited until hardware memory protection was available in
the Intel CPUs while Minux could be booted on an 8088.  Far crappier and
worse hardware - but I think that there were just plain more 8088's floating
around back then than there were 80386's, particularly used ones that
someone had upgraded from.

The FOSS stories that are the most interesting I think is what happened in
the years AFTER these projects got established and started gaining a
userbase.   The story for example of how the warfare between Nvidia and FOSS
finally after decades managed to beat Nvidia and force them to start FOSS
their GPU kernels must certainly be a great one considering Nvidia swore
once they would never allow their chip interfaces to be public and a lot of
reverse engineering was done to write drivers that would work with them.

While the origin stories for FOSS projects are interesting I think most are
meaningless - it's whether the dog started eating when his head was in the
dish is what really mattered.

Ultimately in software we have learned over and over and over again - it's
not how elegant, beautiful, and good it is, it's not how well it works - is
if a lot of people use it or not.  And there are terribly crappy packages
that a lot of people use so "owning" the market is absolutely no indicator
of software quality or efficacy

Ted

-----Original Message-----
From: PLUG <[email protected]> On Behalf Of Russell Senior
Sent: Monday, April 28, 2025 6:50 PM
To: Portland Linux/Unix Group <[email protected]>
Subject: Re: [PLUG] 2001 Linus talk at Computer History Museum

>>>>> "Ted" == Ted Mittelstaedt <[email protected]> writes:

    Ted> Yeah this is just an old defunct religious war bullcrap between
    Ted> Microkernel vs Monolithic kernel, [...]

I watched (or at least listened to) the whole 1.5 hour video and the
microkernel vs monolithic topic was one question out of a couple dozen, and
Linus gave a persuasive answer: that whatever simplicity you think you are
getting with a microkernel gets eaten by the complexity of the
interconnection communication problems. It was a tiny part of the video, so
to boil the whole thing down to "just" seems quite wrong.

    Ted> [...] For all Linus' claims, the Linux kernel today is far, far
    Ted> fatter and has more crap in it than was ever envisioned, [...]

One of his points, which he makes several times, is that there wasn't really
*ever* an "envisioned". He refers to an evolutionary nature of development,
disdain for software design as a preparatory step, and his "broken crystal
ball" when it comes to predicting the future.

Somehow, I want to work in that he mentioned using microemacs[1], which I
once also used, I think in the late 1980s. At some point, when I was still
working in MS-DOS land (pre-1993), I got "Brief". I still have the retail
box for Brief in my attic. When Linux came along, I switched to The One True
Editor.

[1] https://en.wikipedia.org/wiki/MicroEMACS


--
Russell Senior
[email protected]

Reply via email to