Subject: Re: Successful FSBs
From: "Tim O'Reilly" <>
Date: Sun, 22 Sep 2002 09:48:24 -0700

 Sun, 22 Sep 2002 09:48:24 -0700
On 9/19/02 2:08 PM, "Rich Morin" <> wrote:

>>>    O'Reilly (books, conferences, etc.)
>> I'll defer to Tim on this, but my gut feeling is that O'Reilly
>> isn't an FSB, but has a long standing free/open source bent on its
>> many offerings (print/online publication, conferences, etc.).
> What is the essential difference between Tim's publishing a Linux book
> that includes a Linux CD and RedHat publishing a Linux CD distribution
> that includes a Linux book?  When Tim runs a conference (possibly at a
> loss), is this less a part of the community than a USENIX event would be?

Sorry to be responding late to this thread.  I'll start out by saying that I
don't consider myself "an FSB" if an FSB is a binary option. To me, FSB is
not a description of a business's identity, but of some or all of its
business practices, and some or all of its sources of revenue.

It does seem rather silly to me to say that Red Hat is an FSB because they
redistribute free software as part of their business, while O'Reilly is not,
because we don't distribute free software.  RedHat make a great deal of
their money from transactions that are only indirectly related to free
software redistribution.  I haven't studied their financial statements, but
I don't see them redistributing the course materials for their training
classes (which I understand form a very significant revenue stream for
them), and I don't even know how they would redistribute the underlying
value that lets them charge for service contracts and so on.  Of course,
free software is more a central part of Red Hat's ideology and business
promise, and that makes sense to me.

But if FSB means, "free software ideology is a central tenet of the
business,"  I'm even less an FSB.  I believe strongly in both free software
and proprietary software, in free information and in proprietary
information.  I'm always mindful of the remark I heard from Haridas
Chaudhuri, a teacher of Sri Aurobindo's Integral Yoga, when I was a
teenager.  Someone asked him how he reconciled a belief in reincarnation
with the discoveries of genetics, and he smiled and said, "Oh!  You pick the
hat to fit the head."  That's a thought that should be kept handy in the
mental toolchest! Anyone with a one-size-fits-all philosophy is doomed
either to ill-fitting hats or going hatless into the storm, depending on
what's available.

There are a wide range of strategies available for sharing the fruits of
your mind, from free redistribution to managed scarcity.  Which one you
choose depends not only on your objectives, but on your tactics for
achieving them.  (For more on this subject, see the piece I wrote for Nature
a few years back on what Bill Gates and Larry Wall have in common.

But back to the specifics of O'Reilly as an FSB.

* My core mission is orthogonal to the free software/open source debate, but
intensely informed by it.  My vision at O'Reilly is that our business is
about capturing the knowledge of early stage technical innovators and
transferring it to the people who want to follow them.  At heart, we're a
knowledge transfer company.

* As it turns out, many of the innovations in the technical industry have
come from independent developers in a "research" setting, and some of our
biggest opportunities have come from software like Unix and the Internet
that were originally developed and spread without primary reference to the
profit motive.

* What drew us to those technologies was not their openness or lack thereof,
but the fact that they were under-documented, that there was a vast body of
shared knowledge among an early community that had no vehicle of
transmission once the technology was aimed at a wider group.  For example, I
wrote what I believe was the world's first Unix system admin manual in 1983,
when I asked the company I was contracting for the na´ve question, "how are
your customers going to learn about this root stuff?"  I knew the way all of
us at Masscomp learned it was by going and asking Tom Texeira, and I also
knew that customers, at one remove, weren't going to have that option.
(Incidentally, I later reacquired rights to that book, resold them to
Multiflow, where Mike Loukides and later Aeleen Frisch expanded and rewrote
it.  When Multiflow went under, I acquired the rights to the expanded
edition, and it eventually was published as Aeleen Frisch's Essential System
Administration, which has taught hundreds of thousands of people how to do
Unix system admin.)

* Many of the grassroots technologies we document are free software (or
started out as free software), many of them are proprietary, and many of
them are in a gray area in between.  Is HTML free software, for instance?
(I've often argued that "view source" was one of the most important things
about the web, and that it places a big part of the web pragmatically into
the open source world, even if it's ignored by people who are obsessively
focused on software licensing rather than on practical effects.)

* Even when we document proprietary technologies, we're really documenting
the collective knowledge of that technology's best users, not the knowledge
of the technology's proprietary owner.  We scour the net to find people who
seem to know far more about the product than the average, and ask them to
share their knowledge.  And we try to expose the underbelly of the product,
so that people will have as much as possible of the increased power over
their software that's familiar to people in the Unix/free software world.

* We are quite unashamedly willing to say, "our goal is the spread of useful
human knowledge" rather than "our goal is the maximization of the spread of
free software."  And in many cases, free redistribution does not appear to
us to be the best way of achieving our goal.  In other cases it does. When
considering the licenses for books, we sometimes release under free
redistribution licenses if the authors ask us to do so and we think that the
benefits of doing so are greater than the drawbacks.  (See various articles
collected at where I've addressed the various
elements of this equation over the years.)

In the end, I'd say that by the definition of at least the ideologically
inclined in this group, O'Reilly is not an FSB.  But I think we have a model
that a lot of companies that you might consider FSBs, from Red Hat to
Aladdin and SleepyCat, are also adopting.  Like companies that started out
hybrids, such as ActiveState and CollabNet, they are surfing the edge
between free and proprietary in quest of a bigger goal.

Tim O'Reilly @ O'Reilly & Associates, Inc.
1005 Gravenstein Highway North, Sebastopol, CA 95472