Subject: Re: OpenSources "opensourced"
From: "Stephen J. Turnbull" <>
Date: Fri, 21 May 1999 08:49:58 +0900 (JST)

>>>>> "Brian" == Brian Bartholomew <> writes:

    Brian> Stephen J. Turnbull <> writes:
    >> I suspect that "selling software" is a reasonable first
    >> approximation as a description of what Cygnus does.  We just
    >> need to figure out what friction they're using to get people to
    >> buy it

    Brian> You can't download all of Cygnus' software from Cygnus for
    Brian> gratis.  To get a copy from Cygnus you have to buy it.  Nor
    Brian> can you download it from any of their customers that I'm
    Brian> aware of.

For GNUtools, or whatever they're called, the part that most people in
the OSS community care about, you can.  AFAIK Cygnus never withholds
bug fixes, only new features, and those only for a few months.

Agreed, this is a friction, but I don't think it's the one that keeps
Cygnus in that game.

As for ECOS (? the embedded systems stuff), that's a different
ballgame, but as far as I know the software itself is available for
download (Brian, you've been called on the claim that it isn't before;
did you follow up?)  It's the support (and here it's real support, not
mere handholding) that you can't get for free.

Note that since it's embedded, it's not really reasonable to think
that the profits come from "software" entirely, but from bundling with
hardware.  Then the whole "open source" issue is quite moot.  RMS's
strongly held opinion and plausible arguments not withstanding, it is
plausible that embedded systems should be considered widget
components, quasi-hardware, to which the benefits of libre software do
not accrue.  And therefore people who deal in it (according to this
line of thinking) should not be excoriated for not going all the way
to GPL in publically licensing it.

    Brian> Craig Brozefsky <> writes:

    >> "Free doesn't generate software revenue.

Somebody responded (or maybe this is Craig? how are we going to get
OSS textbooks if we can't even get mail attributions right :-)

    >> Free software can generate revenue directly as well.  What are
    >> the folx at Cygnus doing?

    Brian> I feel the 'freedom not price' argument means only 'low
    Brian> price does not imply freedom'.  It does not mean 'high
    Brian> price permits freedom'.  If the price goes high enough the
    Brian> code behaves like proprietary code.

No.  No.  No.  Money is a veil.  As Lao-Tzu would put it, the prices
one can see are not true prices.

The issue is "why can the price go high?"  If it is due to the fact
that there is a market segment which physically can avail itself of
the libre version but prefers the functionally equivalent high-price
version due to frictions in acquiring the software (this is Russ
Nelson's "I got paid $15K because I solved somebody's $50K problem,
[and it's pure profit 'cause I did it with libre softo]" line), it's
still libre.

Or they might be selling support etc, and throwing in the software as
a convenience, then it's still libre but the price is not correctly
attributed to the software qua program.

Or they might be encumbering the program itself in some way
(shareware, nagware, crippleware, vaporware, cookies, bugware, ...).
Then it's NOT libre.  But I don't think this is what Cygnus does.
Modulo a few month's delay, and some frosting in GNUtools etc ---
important yes, but I am not yet convinced it's crucial; I really need
a reliable report on what Cygnus sells for money and what you can
download somewhere, and how they REALLY differ.

Unfortunately the a priori most reliable voices (ex-Cygnus employees
like Ian) don't speak with Brian's boisterous confidence.  But either
way that's an issue of fact, not logic:

Bottom line---yes, Virginia, highly free software can [in theory] carry
(stagger under ;-) a big price tag.

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 two straight lines for?  "Free software rules."