Date: Thu, 03 Sep 1998 06:03:53 -0700 From: "Jordan K. Hubbard" <jkh@time.cdrom.com> To: BSD User Group Hamburg <bsdhh@bsdhh.org> Cc: freebsd-hackers@FreeBSD.ORG Subject: Re: An Open Letter To The FreeBSD Core Team Message-ID: <11246.904827833@time.cdrom.com> In-Reply-To: Your message of "Thu, 03 Sep 1998 12:06:43 %2B0200." <19980903120643.A20537@cons.org>
next in thread | previous in thread | raw e-mail | index | archive | help
> Because of several events happened during the last 6 months we - the
I can appreciate that things can sometimes look worse than they are
(or very different, at the least) from a distance, and I can only say
that neither I nor anyone else in core sees the events of the last 6
months as particularly disastrous or even negative. Yes, the one
"big" event was that we lost a core team member, one who's also hardly
gone away in terms of communications and general offers of assistance
(thank you John), but then anyone who's ever read the FreeBSD 1.1
release notes will quickly realize that personnel turnover in the core
team is hardly a new thing at all. Of the original 13 member core
team present in May of 1994, only 5 are still with us as core members
today. We also have 16 core team members today, showing that
membership numbers have remained more or less constant, a good thing
when you're trying to keep a "management body" at that proper
mid-point between being too large to get any useful work done and too
small to get any useful work done.
The core team of today is also every bit as productive, if not
considerably more so, than the core team of 1993 and if we dug this
project's grave every time a core team member left or was added, we'd
be halfway to China by now! :-)
> We see a linuxization of FreeBSD happen. FreeBSD development
> seems to become event-driven rather than driven by careful
> design.
Then I'm afraid you're somewhat confused about how things have worked
in this project from the very first day of its existence (and having
been here since the very first day, I think I can speak with a certain
authority on this topic :).
We have never had even the slightest desire to see FreeBSD become yet
another academic relic, pursuing some elusive, fuzzy goal of
architectural purity in a rapidly changing marketplace that very
quickly puts such projects in a large pile labeled "interesting
scrap." It's a sad thing indeed, and some such projects of the past
deserved a far kinder fate than they got, but that's just the way this
process works. If software isn't relevant to someone, and hopefully
more than just a couple of someones, its life expectancy is quite
seriously truncated.
This had already come very close to happening to the BSD code base at
Berkeley at least once, and none of us there at the beginning of what
eventually became the FreeBSD project wanted to make the same mistake
twice so we sought to seek a better balance the second time around,
something which preserved BSD's essential cleanliness but was also
responsive to the needs of its user base and actively participated in
the process of encouraging its acceptance (and increasing its very
acceptability).
It's one thing, you see, to produce a pile of interesting bits and say
"here world, feed!" and yet another thing to try and actually become
part of your user base's need-fulfillment loop, actively working on
things that might not personally strike you as all that technically
amazing but nonetheless greatly enhance your product's general
usability for someone who's *not* like yourself. Life is full of
trade-offs like that, not all of them necessarily pretty (or it
wouldn't be a trade-off then, would it? :), and the key is knowing
where to strike the balance. We happen to feel, and a number of
people seem to agree with us, that we've struck a pretty good balance
so far with the FreeBSD project and we're going to keep on going
pretty much the way we have been, modulo any course-corrections that
our users may administer along the way. We have ALWAYS been
user-driven and I would personally rate this as one of the project's
major strengths. Sure, we've made a mistake or two along the way as
well, and if we were absolutely perfect in our judgment at all times
then I doubt we'd be doing free software at all - we'd be making a
fortune investing in the stock market. :-)
In summary, you can't please everyone all the time and clearly we
haven't entirely pleased some members of the Hamburg FreeBSD User's
group, but I think we're pleasing the majority of our users and that's
all that really counts. You can't make everyone happy, but if you can
make *most* people happy then you're doing a better job than several
major world religions. :)
> While we believe that a very good technical background is
> required for being a core team member, we feel that communicational
> and organizational management skills are a conditio sine qua non.
A quick word about core team members here: If someone ever invents the
perfect core team member, I'll be first in line for as many as they
let me take away, but in the meanwhile I have to live in the real
world and the real world poses some significant constraints. Some of
those constraints for a core member is that they have to be someone
who:
1. Is willing to spend a significant amount of time, without
pay or even thanks, working on the project on an ongoing
basis and not quitting the minute somebody angers them or
things get difficult.
2. Is willing to potentially do many things for FreeBSD that
they probably wouldn't actually do for money, like stand in
front of hundreds of strangers and give presentations or
work for hours on a driver for some obscure card they don't
even *like* or want, but some number of users really need
support for.
3. Is technically proficient enough to understand most of what's
going on most of the time and be able to comment
semi-authoritatively on almost any aspect of FreeBSD when
cornered in the hallway by some FreeBSD fan who's got a
million questions.
4. Can be said, when all else fails, to truly have FreeBSD's best
interests at heart, even when they're no longer personally
involved with it. I think it's telling that the great majority
of ex-core members still care deeply about FreeBSD and continue
to help out from time to time. There have been very few "messy
divorces", even ones which might have seemed so at the time.
5. Has a certain "je ne sais qua", but I don't know what it is.
[rim shot :)]
Finding someone who embodies all those qualities, especially #1, is
difficult as it is without finding people who are also models of
decorum and emotional stability. In fact, there's something about
free software which seems to attract a lot of people who define
"stable" as "didn't wake up in bed with an axe again this morning" :-)
Speaking as a general rule, it's also a really bad idea to judge the
project by the conduct of any one of its members. We all have our
good days and bad days and, as I pointed out before, the project
itself goes on, frequently under the guidance of different people as
the core team turnover stats I cited earlier also prove. What's the
point of getting all upset over the antics of somebody that may be no
more than a footnote in the handbook in another year's time? Just so
long as the project itself is being well taken care of and there is a
constant supply of fresh and enthusiastic new people stepping in to
take up the slack, you're probably doing as well as any free software
project can.
Let's not lose context here folks - this is free software and you
simply don't have all the nifty artificial "enticements" of a steady
salary and other factors which keep commercial software groups
together. It's pure fallacy to directly compare commercial
development organizations with their analogous free software groups,
even those who are starting to make money as individual members, and
you have to adjust your standards just a bit to compensate for the
differences in operating model or forever suffer from perceptual
problems about just what you should and should not expect. It's
hardly a *worse* way of operating, and I'd say that the FreeBSD
project is doing exceedingly well with its operational model at the
moment, it's just a different one.
> Almost no substantial technical discussion takes place anymore
> on -hackers; some of the core team members doing architectural
That's probably more a product of our success than anything else, I
hate to say. The erosion of forums like this, frequently accompanied
by various people pining publically for "the good old days", is a
simple fact of life as the users flood in and begin dramatically
affecting the signal-to-noise ratio - I've seen it in the Linux groups
for some time and now I'm starting to see it here. Oh, I do what I
can with my little form letters which redirect users to the right
mailing lists, but it's really just an attempt to hold back the tide
at this point. We're still growing, a process which is nowhere close
to finished yet, and the influx of new and admittedly often lazy users
who just post to the first forum they come to ("charters? what
charters? Handbook? Huh?!") is only going to increase and drive more
people out of the various "tech" lists. I regret this as much as
anyone, of course, but short of going to a very "expensive" moderation
scheme or something yet to be invented, I don't see many alternatives.
We can either work on FreeBSD or we can read mail - which would most
of you prefer? :-)
> We ask the core team to reconsider what they think the purpose of the
> core team is, and what the best way is to achieve it.
Having discussed your mail and considered it as you request, I think
we really have to say that we're quite pleased with how things have
been going for FreeBSD these last couple of years, it certainly being
time of many significant gains across a very wide front (including
many developments of interest which none of us could have
anticipated). All of us are also more than ready to admit that things
have not gone absolutely perfectly the whole time, and in many ways
we're sort of learning this as we go along (not many good books on
starting your own free OS project :-), but I think things have gone
very well indeed nonetheless - certainly better than most of us ever
expected.
In any case, since there are obviously going to be many sides to an
issue like this, it probably wouldn't help the already low SNR of
groups like -hackers (which was one of the complaints raised, after
all) to debate this back and forth ad-infinitum. About the best you
could hope for is that some number will agree that there are problems,
some number will feel that things are perfectly fine, both sides will
feel very strongly about their point of view and, in the process of
proving this, we'll generate a lot of heat with probably little (or
no) light which would be of actual use to anyone.
- Jordan (speaking, by arrangement, for the freebsd core team).
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-hackers" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?11246.904827833>