Subject: Re: funding indirect services
From: "Stephen J. Turnbull" <>
Date: Fri, 28 Jul 2000 11:15:26 +0900 (JST)

>>>>> "Crispin" == Crispin Cowan <> writes:

    Crispin> The Open Source Revolution[tm] has essentially
    Crispin> demonstrated that the way of doing this is that people do
    Crispin> it for free, for themselves, on their spare time, and
    Crispin> distribute the results.

    Crispin> "Stephen J. Turnbull" wrote:

    >> You're taking history too seriously. [...]

    Crispin> Contest if you like, but that's really how it happened.
    Crispin> Case in point: Apache.  Built by a bunch of web masters
    Crispin> who started contributing patches to the NCSA web server,
    Crispin> hence the old gag "a patchy server", which begot the name
    Crispin> Apache.

I don't contest the history.  I do say that "essentially demonstrated"
is your way of saying "post hoc ergo propter hoc," a fallacy.  It may
be the only way that so far has happened on a large scale (but I think
<DrumRoll> The Internet's software infrastructure's history proves
otherwise), but it does not preclude other ways, including FSBs, of
accomplishing development of large-scale open source software.

    Crispin> Wanting to get paid for doing this kind of thing is
    Crispin> "shareware"; the ancestor to Open Source software.

    >> I contest your geneology here.

I meant I don't think shareware is the ancestor of Open Source.  I
agree with Richard Stallman[1], not with my Scottish economist
intellectual ancestry: for programmers, the impulse to share comes
first, the impulse to profit (a close ;) second.

    >> Also your history, dunno ee from the ancestor of cs but
    >> imagemagick and xanim have been around for quite a while.  I
    >> don't think they were spawned by xv.  They have their own
    >> raisons d'etre.

    Crispin> xanim is an MPEG viewer.  XV and EE are
    Crispin> GIF/JPEG/TIF/etc. viewers.  Different animals.

Well, xanim does GIFs.  I misremembered that it did other stuff (eg,
JPEG) - it doesn't.  Still, there have been a lot of viewers around
for a long time.

The point I'd hoped to make about xanim was precisely that its primary
purpose was not a free version of xv.  There are a lot of motivations
for writing any given free software app, and not all of them involve
removing undesirable restrictions from existing app categories.

    >> And there still seems to be a heck of a lot of shareware out in
    >> the Windows and Mac worlds.  But there never was all that much
    >> in Unix.

    Crispin> Shareware plays better on the desktop, where large
    Crispin> numbers of relatively non-technical people consume goods,
    Crispin> and are willing to pay (modestly) for them.  In the
    Crispin> server/infrastructure world, when a technically savy
    Crispin> person is presented with some utility code that does
    Crispin> something useful, but is not unimaginably difficult to
    Crispin> write, then they are just as likely to re-write it as
    Crispin> Open Source as they are to pay the shareware fee.

In fact, if sourceXchange and can get rolling, they may
be more likely to rewrite it as open source _to make a couple bucks_
than to _pay_ the shareware fee.

I know of at least one project (a GTK+ version of XEmacs) that has
been mooted for several years but actually happened only because of a
patron listing on sourceXchange.

    Crispin> So the "business" way to pay the tab for this is to be
    Crispin> employed by a company that would materially benefit from
    Crispin> the way in which your widget makes things better.

    >> But I think Rich's point is that it pays even better if you own
    >> the company.

    Crispin> Here's the core point of contention: I claim that the
    Crispin> open source "potlach" effect does not scale well.
    Crispin> Individuals contribute to open source projects because it
    Crispin> is in their personal interest to do so, because the
    Crispin> combined better product makes their lives easier.

Again post hoc.  I just gave an example where that's not true.  The
reason that "potlatch" has been the primary driver (and again the
history of the Internet is a killer counterexample IMO) of OSS
development in the past is the transaction cost of finding,
negotiating, and supervising contracts where neither party will own
the product.

OSS licensing inherently limits revenues, and therefore transactions
costs loom larger.  But this is precisely why I have hope that markets
in OSS development will bloom:  we have the technology to reduce those
costs, people are implementing it.  This is exactly a field for
"ethical" business.  People are looking to make money, in substantial
chunks---in the XEmacs/GTK project, K$15 for 4 months of spare time,
doing something he wanted to do anyway---under the constraint that the
product be contributed to the community.

That constraint doesn't mean that some programmers[2] wouldn't hold
out for K$30 if they thought they could get it.  But without these
markets, the patron and the programmer would never meet.  And that's
what you are observing historically.

    Crispin> This effect does *not* scale up well: it is not easy to
    Crispin> imagine how a company could profit from this.  You claim

Ask Brian and Bernie.  You may find that others' imaginations are more
powerful than yours in this dimension.  :-)

    Crispin> that the non-existance of companies doing the same thing
    Crispin> as the individuals belies the hypothesized effect.  I

Read me again.  I don't claim that.  I claim that the future is not
bound to look like the past.  I don't claim the future will look
nothing like the past; history is not irrelevant.  But we're here
looking for ways to change the world ... and at least some of us to
find new ways to make money opened up by those changes.

    Crispin> claim that it is a small-scale phenomena that does not
    Crispin> scale up well.  Kinda like the way there are no 900-lb
    Crispin> water beatles :-)

Except for the tragic loss of Lennon, the Silver Beatles probably
would weigh about 900 lbs by now.  McCartney and Starr certainly have
been scaling up.  :-)

Markets are a small-scale phenomenon whose biggest advantage is that
they do scale, in fact, the bigger the more efficient.  I don't know
if the markets we see today at sourceXchange and are the
right institutions yet, but once the formula is refined, they will
scale.  (Out-on-a-limb type prediction, not claiming "scientific
truth" here.  "You heard it here first."  :-)

[1]  No, it's not the first time.  ;-)

[2]  I have no idea whether Bill tried to negotiate compensation, and
it's not relevant.  Surely some people will try and many of them succeed.

University of Tsukuba                Tennodai 1-1-1 Tsukuba 305-8573 JAPAN
Institute of Policy and Planning Sciences       Tel/fax: +81 (298) 53-5091
_________________  _________________  _________________  _________________
What are those straight lines for?  "XEmacs rules."