Subject: History of the NetBSD Foundation
To: None <email@example.com>
From: Charles M. Hannum <mycroft@MIT.EDU>
Date: 09/01/2006 14:29:07
I don't really like to focus entirely on the past, but many people have
inquired about the seemingly odd history of The NetBSD Foundation. So
here it is.
When we started the NetBSD Project, it was a loose organization of
people around the world. The servers were initially supplied by Chris
Demetriou and hosted at Berkeley, though it became clear that this
arrangement was not going to last. Chris was busy trying to graduate
and did not have time to keep things running. (As I've said before,
all of the open source OSes -- including NetBSD -- were buggy at that
time, and some hand-holding was required.) So, still on Chris's
equipment, much of the services actually moved to the Free Software
Foundation office at MIT.
This arrangement was reasonable at first. It allowed me to make good
progress on fixing some of the important problems. I also institued
the switch to remote CVS, and to encrypted logins. Access to CVS was
much more reliable, and people were able to get more work done.
But this situation could not work long term. First of all, we needed to
upgrade equipment; we could not keep running on the same hardware
forever. We also needed more bandwidth to feed an increasing number of
users. Both of these things required resources that I could not supply
myself. In addition, we saw a need to consolidate copyrights, both to
make it easier on people using our code, and to make it practical to
enforce them if necessary. So, we decided to create the NetBSD
Foundation, to manage copyrights and launder money into servers and
I may be misremembering the timeline, but I think it was at the summer
Usenix conference in 1994 that we actually got together in person and
signed the paperwork. We (and by this I mean primarily John Conklin)
had drafted a basic set of "for any legal purpose of corporation" type
bylaws and articles, and we did. (BTW, Usenix conferences were often
the only time we met in person.)
I want to point out specifically that the Foundation had only a board
and an executive staff at this time. It was not a membership
organization, nor did we advertise it as such. Day-to-day stuff was
still run by the "core group" and our volunteer sysadmins, both groups
predating the Foundation and operating more or less independently of
it. We did start having people sign the infamous "developers agreement"
in order to get commit access to our CVS repository. This was merely a
legal shield for the Foundation, in the same way that the corresponding
agreement used by CSRG (Berkeley) was. It did not confer any
"membership" status, nor was it intended to; I should know, because I
Things plodded along. We got a generous donation of a server from
iki.fi (instigated primarily by Tatu Ylonen, when I followed up on an
offer he had made some time before), and colo space at hut.fi to run it.
Part of the reason for pursuing this was to move the source code outside
the U.S., because of the restrictions on crypto developemtn at the time.
Other colo arrangements at NASA, PAIX (with the support of ISC), ISC's
own space, 200 Paul, etc. followed. In general there have been problems
with all of these arrangements, but they have been "good enough" to keep
the project running.
Around 1998, it became much clearer that not having a reliable CD
distribution was a problem. Also, some other projects (notably FreeBSD)
were largely funding themselves through CD sales. So, along with Herb
Peyerl, I started what we called "The NetBSD CD Project" -- a separate
entity from the Foundation, so that the finances were not mixed.
Shortly after, I also started "The NetBSD Mission" (registered in
Cambridge, MA) to sell other merchandise, primarily T-shirts. The real
intent here was to help popularize NetBSD through promotion; a secondary
goal was to eventually turn enough of a profit so that I could focus on
NetBSD and not have a day job. For various reasons, this didn't work
In 1999, I also started hitting the conference circuit -- Comdex,
Usenix, ALS, and many others -- running booths and demos of NetBSD.
This required an enormous amount of time (even just making the system
"demoable") and money (primarily funded by my own consulting). Most of
it was done under the "NetBSD Mission" name, because the NetBSD
Foundation had reached a sort of stalemate -- due to attrition, there
were only two board members left. This work was very successful in
stirring up interest in NetBSD, and getting some companies to take
NetBSD seriously and use it in products.
What I also learned in this time was that the secretary and treasurer of
the Foundation, Christos Zoulas, had not actually been doing the
paperwork, and the Foundation had fallen out of good standing. Contrary
to some statements, this does not mean it simply ceased to exist -- it
still had assets, and there is a clear ownership of such assets. I want
to stress that I was not aware of this; every time I inquired, I was
told that it was being taken care of. This does not absolve me of
responsibility, though -- clearly I should have been getting copies of
the paperwork, and I did not.
This was turned into a political fiasco, primarily by Perry Metzer, a
name that some people will recognize and probably cringe at for other
reasons. It was conflated with issues about turning the Foundation into
a 501(c)3, etc. I was made the scapegoat, even though the actual lack
of compliance was largely the secretary and treasurer's doing. A new
set of bylaws was drafted -- not by the existing Foundation, by a friend
of Perry's who was never trained in IP or non-profit law -- and a "vote"
was held. It was ignore that this "vote" had no legal standing because
the people "voting" had no legal standing WRT the Foundation.
Nevertheless, the resulting changing of bylaws and board were filed by
Christos, without the approval of the existing board. That act was
This is particularly bothersome because there is no good reason for it.
I had attempted several times to get Christos to meet and work out the
problems with the bylaws (one of which is that they meet neither the
spirit nor the letter of the law WRT a non-profit), but he simply would
not do it.
This issue gets muddled a lot for a number of reasons. The new bylaws
*do* have a "membership" function -- this being important to have a
voting mechanism. They also stipulate one of the 501(c)3 requirements.
However, there are other problems that still need to be fixed. I was
not against the bylaws in general, but I felt strongly that the errors
needed to be fixed before they were ratified.
Over the course of this, there has also been an issue with the
stewardship of the original TNF paperwork. When our original secretary
vanished, we had considerable difficulty getting the paperwork from him.
Eventually it was transferred to someone else we had asked to do it, but
then we had difficulty getting the paperwork from him. Today, the
paperwork is in my possession, for safekeeping, until the issues with
the Foundation's governance are resolved. I have, however, provided an
exhaustive list of the "developers agreements", including when they were
signed and received.
That's the Foundation's history in a nutshell. The shell does indeed
contain a lot of nuts.