Subject: Re: Successful FSBs
From: "Tim O'Reilly" <>
Date: Sun, 27 Oct 2002 08:43:11 -0800

On 10/27/02 12:41 AM, "Stephen J. Turnbull" <> wrote:

> But innovation requires matching invention to market.  If necessary,
> create the market: hire suits, educate the potential customers, both
> by advertising and by going to shows and doing training sessions for
> your product.  Make slick demos.  But that costs money and requires
> giving up some control (at least) to the marketing department, and is
> alien to the developer mentality.

Actually, I think that there's a natural progression here.  I've been doing
a talk lately called "Watching the Alpha Geeks" that explores this thought.
Hackers push the envelope because they can; entrepreneurs come along when
the market is a little riper, and start to make businesses; eventually, the
innovations are built into the next level platform by a dominant player.

The problem I've been trying to work is to get the free software community
to see this progression, and where folks like MS are taking some of the
ideas that free/open source software has pioneered, so that we aren't
playing catch up again with the platform we ought to have built first.
> Until we learn to partner with suits, FS innovation is going to be a
> hit and miss thing.

I don't think that FS innovation is absent.  It just isn't celebrated
enough.  And it isn't followed through enough.  The follow through and
commercialization often comes from non-free businesses who cherry pick the
ideas and take them to the next level.  It's not the innovation that's hit
or miss, but the commercialization of that innovation.

>  And by "we" I mean the whole community.  The
> "arch" whinge-fest demonstrated that there are a _lot_ of people on
> this list who get it, and (funnily enough) I bet members of this list
> account for far more than than their share of FS innovations that have
> made it to general use.  I imagine most FSBers have made their peace
> with their marketing departments (or the need to do marketing
> themselves).  So as in my reply to Tim, I'm thinking about the
> relationship of non-B FS-organizations (like XEmacs, GNU, the 10000
> noncommercial projects hosted on Sourceforge) to not-necessarily-FS
> B-organizations.  Sitting at one interface are GNU, SourceForge, and
> Bitkeeper: an interesting case study in contrast, it strikes me.  Then
> there are and sourcexchange (RIP) at another interface.

People keep saying sourceExchange (RIP).  But is alive and well,
and doing a lot of this interfacing.,,,, are all very successful
projects managed by
> Unfortunately, there's a substantial school that's hostile to that
> whole train of thought.  Business is about withholding the product
> until you're confident of being paid.  Free software, especially
> copyleft, is about getting the product out there, and making sure it
> stays out there, whatever downstream entities want to do with it.  It
> takes a sense of balance and compromise to work out the apparent paradox.

I see free software as a great way to deliver what Clayton Christenson
called disruptive technologies.  They don't have to be commercialized
immediately for them to have a huge impact.  And in fact, they are often not
very successful at first.  Look at how the personal computer developed, from
early hackers through pioneering entrepreneurs like Apple through IBM's
brilliant decision to "open" the design of their PC, leading the way to
commodity hardware, and ultimately the wintel platform.

Note, however, that the effect of commodity hardware was not to make
hardware free but to make hardware cheaper, and more standardized.  It had
all kinds of secondary effects which pushed proprietary IP higher up the
value chain.  Very specifically, it was commodity hardware for the PC that
led to a major shift of power from hardware to software vendors, a paradigm
shift that was exploited by Microsoft, the luckiest and most-prescient of
the software entrepreneurs.

Similarly, putting software out under the GPL just means that a lot of the
commercial value will accrue to people who build proprietary IP on top of
it.  RMS *thought* he had a way to keep people from building proprietary
software on top of his free software, but he didn't see far enough ahead to
the paradigm shift to network computing, which would allow a google, say, to
build on top of linux, and keep all its higher layers proprietary.  Because
google isn't distributed, the GPL never applies.  (And yes, I know Richard
is working on this "problem" now -- I believe I was the person who put the
bug in his ear about it, when we both spoke in Berlin at the Wizards of OS
conference in 1999.  I know he wasn't thinking about it before then because
in our back and forth in the Q&A after my talk, he said the issue wasn't
important!  See or the
more granular links from

I've been saying for years that the shift towards commodity software
(whether free or just open standards) would lead towards a new paradigm in
which money was increasingly made on services.  (At one point I was calling
it infoware, now I'm saying web services and 'the internet operating
system', but the point is similar.  People don't pay for the software, but
for the services the software delivers.)

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