GNU Hurd port

John Moser john.r.moser at gmail.com
Wed Dec 9 18:38:18 GMT 2009


On Wed, Dec 9, 2009 at 11:17 AM, Patrick Goetz <pgoetz at mail.utexas.edu> wrote:
>> Subject: Re: Supporting a GNU Hurd port?
>> From: Scott James Remnant <scott at ubuntu.com>
>> Date: Tue, 08 Dec 2009 17:14:38 +0000
>> To: Danny Piccirillo <danny.piccirillo at ubuntu.com>
>>
>>
>> Speaking as the guy who maintains the boot and plumbing layer, I am
>> completely and utterly uninterested in such a port.
>>
>
> My question is why are we even talking about this?  The linux kernel is
> by far the most robust, dynamic part of the whole distribution, and it
> should be quite clear that there aren't enough developers to maintain
> even the software that's in the distro now.

I'm the guy that thought street lamps run on burning gas was a
perfectly workable idea, but using some sort of glowing electric wire
without fire would be even better.

Would you have rather kept your indoor lighting as kerosene torches on
the wall?  At the very least you'd be warmer in the winter....

A Ford Shelby V8 still has a live axle in the back.  It works, but an
independent rear suspension engineered properly (the one in the Cobra
03/04 was crap, the LSD was done wrong, and it got horrible wheel hop;
it was basically a FWD non-diff suspension for a 200ftlb car shoved
into a RWD car with a rear diff getting 400ftlb torque) would be much,
much better and the car would be able to corner VERY hard without
flipping or skidding sideways.

At some point, all of these things were non-obvious; at some other,
more recent point, they were considered "interesting, but not
important."  These days, we're sitting around laughing at Ford because
the Mustang has LOL LIVE AXLE 1920 SUSPENSION....

Right now, a microkernel is still between "non-obvious" and
"interesting, but not important;" we haven't quite determined if there
is a real benefit, we'll argue over if there is or isn't, and most
people who decide there is a significant benefit declare that it's
simply not significant enough to matter even though they've never seen
it put into full practice.

We could be "innovative" or "technological leaders," but that puts us
all at risk of "doing something stupid and having it not work out
quite right."  It's more palatable for most people to just sit back,
wait for someone else to do it, and then either watch it collapse or
play catch-up; who the hell wants to take risks on new technology
themselves?  Especially if what, maybe we use HURD and find out Minix
was much better and surpassed HURD support-wise; or we use Minix and
find out HURD was just better in a HUGE way.  Maybe a third player
will come up and omg, The Perfect Kernel!

The core of the whole argument is, of course, that some of us see
these things as more interesting than others.  Scott simply doesn't
want to do more work -- trust me it's a hell of a lot of work, and
I've already given the best way to attack it ("best" being "mitigates
any extra work in the future"), and that's still a hell of a lot of
work.  Others here don't trust the technology, still others are
(mistakenly) convinced it's already been determined that the
technology is a worthless research project with no practical benefit.

And then there's the crowd that believes this is the future and we
need to get on it RIGHT NOW and doesn't understand the concept of
prioritization.  Our pet projects obviously deserve 100% of the
resources, right?  Ubuntu should go to Minix right now (I am a Minix
fan), they should implement Linux d-bus and NF_NETLINK and udev stuff
so the boot process is drop-in, they should start porting FreeBSD
driver code to the Minix code base, and not a damn thing else matters.
 That makes sense, right?

No?

It probably doesn't make sense if you have any freaking clue about how
a large project is run; but why should reality get in the way of your
pet projects?



More information about the Ubuntu-devel-discuss mailing list