Re: design goals vs mechanisms

[Marcus Brinkmann]

At Thu, 27 Oct 2005 17:06:27 +0200,
Alfred M Szmidt wrote:
> 
>    When I came to the Hurd in 1997, it was essentially dead.  It was
>    not yet officially declared death, but it was pretty close.
> 
> You must have a skewed view of history, the Hurd was no where near
> `essentially dead' in 1997.

Huh, yeah, it should have been 1998.

> on FOO.  During the past 5 years the Hurd has seen about as many
> changes that it saw during 1996.  For the record, you didn't
> contribute a single line until 1999.

I joined the Hurd in April 1998, not 1997 as I wrongly wrote above.
My first post was, ironically, an email to help-hurd with the subject
title:

Subject: hello, seems that hurd is stuck.
Message-ID: <address@hidden>

I am not going to argue with you about the meaning of the word "dead",
but GNU 0.2 was released in June 1997, and was so bug-riddled that you
could hardly do anything at all with it except to look at the blinking
cursor.  I was the first one to report the critical bugs in it, one
year after the release, and I helped debugging them.  The mailing
lists were extremely low traffic, until I popped up.  You can ask
Thomas and Roland themselves if they would have made so many changes
to the Hurd tree in late 1998 and 1999 (and later) if I (and others)
hadn't been there to poke them.  There has been significant
development independent of mine from Mark Kettenis and others, but at
some point even that simply halted.  In fact, if I look back into the
archives, then beside Thomas and Roland I may be the only one left
from 1998 who is actually still active on a consistent basis (even if
it is not in an area you appreciate).

Note that the changelog entries numbers you quote support this, as
well as the fact that Thomas (and Roland I think) stopped working for
the FSF in 1998.

And if all this is not enough, I _stopped_ doing consistent and active
upstream development on the Hurd sources on Mach in 2003, just when
the number of changelog entries drops dramatically.

You may also want to check out what the plans of the FSF with the Hurd
were around about 2000/2001.  There was a time where the Hurd
apparently was close to being stomped into the ground for good,
officially.  I do not remember what made us panic, but there was an
announcement or a quote from a public speech RMS gave that seemed to
prepare the ground for declaring Linux the official kernel of the GNU
system.  If I remember it correctly, it was partly due to Neals and
mine intervention (at the LSM in Bordeaux 2001) that it even got
another chance (although there may have been other deciding factors).
One night, we gave a quick demonstration of the Hurd to Bradley Kuhn,
and asked him for more time.  At that point we _did_ believe that we
only need to fix a couple of bugs to put it into a releasable shape.

This was the basis for the eventual "announcement" (in a Q&A session
at a speech in India) by RMS that the Hurd could be released in 2002
(see for example http://slashdot.org/article.pl?sid=02/03/12/1236243).
In hindsight, this was simply embarrassing.  Actually, it was
embarrassing at the time it happened, because by 2002 we were not
quite as naive anymore as the year before.

Thanks,
Marcus