Subject: Re: Successful FSBs
From: "Tim O'Reilly" <>
Date: Tue, 24 Sep 2002 13:07:08 -0700

On 9/24/02 4:52 AM, "Stephen J. Turnbull" <> wrote:

stuff about O'Reilly snipped.  Thanks, Stephen, for the nice words.

> How do we justify calling that an FSB in a way that's useful to
> startup FSBs?  I don't have a good answer, although I'd like to have
> one.  Maybe you have a generic model in mind (ie, lines of business
> that are as profitable for free software as they are for nonfree
> software), or perhaps there is something specific about ORA
> publications and support services that is enabled by free software?

Well, I wouldn't call us an FSB.  But I'm also not sure that making up a
rigid definition is all that useful.

I've always liked to point out that humans like to think in terms of
boundaries, but that natural events tend more to gradients surrounding an
increasingly dense core.  Where is the edge of the earth's gravity?

I suppose, in practice, you can define a pragmatic edge, but it's always
changing, just like the shoreline between land and sea.

I think it's more interesting to ask the question "What business practices
associated with free software are the most successful?" rather than defining
a business as an "fsb" depending on whether or not it meets some litmus

My guess is that very few, if any, successful businesses will be rigid FSBs
by the hardline definition of this group.  Every single example I can think
of has a mix of free and proprietary in its business model.  I just don't
think that drawing a bright line makes a lot of sense.

Earlier in this thread, someone mentioned that I like to cite ISPs as
examples of businesses that profit from free software.  I challenge you to
distinguish them from even the early Red Hat except on grounds of ideology,
and the fact that they didn't actually distribute software in the 1980's
paradigm, but instead delivered it in the emerging paradigm of the 21st
century.  Very little proprietary software for many of them; in some cases,
major contributions back to their community. (I like to point out that Uunet
was driven originally by usenet, and that Rick Adams was the author of B

Here, to me, are some interesting questions:

* To what extent is the software sold or "performed" by the business shared
by a community of developers rather than controlled by a single entity?  Do
those developers compete or cooperate in extending it?  (I love the fact
that Apache was started by a community of users.  This is far more
interesting to me than the Red Hat branded ketsup model.)

* What circumstances control the different economic returns from a strategy
of maximizing usage vs. maximizing revenue?  That is, when is it better to
give your software (or documentation, or music, or whatever) away in order
to get users and visibility? When is it better to restrict distribution or
user license rights in order to get revenue?

* What kinds of products and projects are not created by the natural
dynamics of free software communities and benefit from added incentives
created by managed scarcity (i.e. Redistribution restrictions)?

* What are the useful competitive tools that matter when the product is a
freely redistributable commodity?

* What are the "real" ownership practices of free software communities,
independent of licenses.  (See my earlier note about RMS and free
documentation.)  Eric Raymond addressed this in "Homesteading the
Noosphere."  I agree that there's a strong system of moral rights that
amounts to "property rights" that play a bigger role in this discussion than
is often acknowledged.

* What are the possible revenue streams associated with free software?  How
hard are they to achieve? (i.e. What are the costs beyond the development of
the software required to realize them, versus the possible return?)

I think that answering these questions is the point of this discussion, but
the framing of the question around "identity" ("What is an FSB?) and narrow
definitions seems to me to get in the way of the answers.

If you want a definition, I'd say something very broad, like "An FSB is any
business that uses free software as a significant part of its business
strategy, and that profits, directly or indirectly, from the wider use of
free software."

And yes, that includes O'Reilly, and IBM, and Sun, and ActiveState, and
CollabNet, as well as Red Hat.  But if you don't use a very broad
definition, I think that before long you won't have any members left in the
set at all.  Red Hat, and SleepyCat, and so on, all have at least some
proprietary component to their business -- often more than outsiders realize
-- and the distinctions that are made seem to me to be largely of academic
interest.  If Red Hat gets 25% (numbers pulled out of a red hat) of their
revenue from proprietary training, and over time they add other proprietary
products and services, at what point do they cross over the line and
suddenly stop being an FSB?

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