From mboxrd@z Thu Jan  1 00:00:00 1970
Message-Id: <200407280408.i6S483P10762@augusta.math.psu.edu>
To: Fans of the OS Plan 9 from Bell Labs <9fans@cse.psu.edu>
Subject: Re: [9fans] (no subject)
In-Reply-To: Your message of "Tue, 27 Jul 2004 11:54:16 BST."
	<755cce86156cfbde0dee14b5470d0ba5@quintile.net>
Date: Tue, 27 Jul 2004 21:08:03 -0700
From: Dan Cross <cross@math.psu.edu>
Topicbox-Message-UUID: cea6c332-eacd-11e9-9e20-41e7f4b1d025

"Steve Simon" <steve@quintile.net> writes:
>
> [...]
>
> I was disturbed by David and Russ leaving the labs, and I am
> worried that the 9fans community is losing momentum - perhaps
> there are many great projects under way but I don't hear of them.

It is (losing momentum).  But then, it only had very little to begin
with.  The problem, as I see it (several months removed from the
`scene' as I am right now; I was probably lobbing hand gernades over a
berm somewhere when they left the labs), is that the driving force of
recent years is gone: building a comfortable system for doing research,
but one that's interesting in its own right.

With the deflating of Murray Hill as a hub for activity - mainly
development, but also pushing on the license issues and such - there's
little direction in the system.  Everyone who was influential in the
early days was using Plan 9 for one of two reasons: they wanted a nice
environment to do computer science research, or they were interested in
operating systems research and Plan 9 was an interesting thing to
study.  But that's the kind of thing you can do when you work in a
university or an industrial research lab that starts projects with the
understanding that it might be 30 years before they show a positive
ROA.

Most of those people now has other things they have to do; they can't
afford to continue hammering on Plan 9 full time anymore, and no one
single organization seems to exist to fill the void.  What's more, as
they leave the `community' (btw- I really hate that term), or at least
become more reclusive, they take with them the sense of taste and style
that made Plan 9 unique.  After all, the thing that makes the system
unique is the aggressive development of ideas taken from earlier
systems, with some new stuff thrown in.  But the thing that makes it
*good* is the quality of the system as a whole.  Anybody can probably
re-implement the ideas in Plan 9, but the trick is to do it well, and
that is very much harder indeed.

I doubt the appropriate environment to take up the torch of Plan 9
development exists anywhere in the world: a bunch of really smart
people that also have good style sitting around (by agreement!) in the
same physical room with the capital and support to work on whatever
they want, and who want to work on something like this.  Much as people
want to tote the ``feel-good'' party line of open source and disjoint
development efforts coordinated over the Internet, I really doubt
that's the answer: the world has yet to see a *good* project emerge
from such techniques.  What's lacking is a suitable research center to
pick up where Bell Labs left off and push the system further.  Without
systems research, you can't sustain a research system.  There's no
systems research going on, therefore the system is withering and
dying.  Without good people driving it, the system will take a nose
dive because, frankly, most people aren't skilled enough to do it
properly (look at Linux!).  If Rob Pike is a master chef, then most of
us are Subway sandwich artists, and an even greater portion of the
population can't even make peanut butter and jelly sandwiches.

Or maybe I'm just overly pessimistic.  As Geoff points out, people are
using it and satisfied with it, and maybe that's enough.

> It has been said before, the limited of support plan9 gets
> in the world is partly due to its steep learning curve, and
> partly the lack of "standard" applications.

Yes, but bear in mind that Plan 9 wasn't really designed with world
domination in mind.  It's a research system.  That it evolved into a
very comfortable environment for the people who built it to do *other*
research on made it possible to pursue other goals on it is almost an
accident of history.  So the question is, who cares how much support it
gets in the world?  Should the goal of any system apriori be to attract
lots of users?  I don't think so.  It should do something interesting
in a clean, elegant way.  But that's just my personal opinion.

> Writing a web browser difficult, we all agree, however editing the
> wiki to make it more up to date, accurate and helpfull is easy
> (and yes I have been).

That's a noble and worthy goal in its own right, and I won't
argue with it.

	- Dan C.