Subject: Re: How accurate is Metcalfe's law? (Was: Ximian software)
From: "Stephen J. Turnbull" <>
Date: 10 Jan 2002 12:28:39 +0900

>>>>> "Ian" == Ian Lance Taylor <> writes:

    Ian> "Stephen J. Turnbull" <> writes:

    >> >>>>> "Ian" == Ian Lance Taylor <> writes:

    Ian> There are some great ideas which have been extensively mined:

    >> You've got the wrong kind of greatness in mind.  Those are
    >> great inventions, creating _potential_ social value.  Putting
    >> those inventions at the service of users (what economists call
    >> "innovation") is a different kind of idea which programmers
    >> give short shrift.

    Ian> First I say that I don't think there are diamonds in
    Ian> programming.  Then you say, what about Unix?  Then I say,
    Ian> yes, there are few diamonds, and Unix is one of them.  Then
    Ian> you say, no, that's not what I meant.

    Ian> So I don't know what you mean.

Well, I haven't been very clear.  Unix _is_ a diamond in the sense I
meant.  The fact that it solves certain system organization problems
in an elegant way is nice, but lots of systems are elegant.  (Anybody
here write production code on Turing machines?)  What is important is
that somehow between the technical elegance and AT&T marketing and the
FS/developer sharing ethic it fostered, Unix has become the standard
that you have to beat.  Because users use it.

    Ian> Putting an invention at the service of users is exactly what
    Ian> I intended to do when I wrote my UUCP code,

I once wondered, but, hey, "wow! that was you?!"  :-)

    Ian> I have a hard time calling that a diamond of any sort.

I don't.  I'm impressed.  (But see?  "There you go again," deprecating
what was arguably a big contribution because technically it was
"obvious" to you, and just involved "hard work" to implement.)

    Ian> It was neither invention nor innovation in the sense in which
    Ian> those words are used by ordinary people.

Because ordinary people use them interchangably, to mean the idea
itself.  They don't have a word for what economists call "innovation"
(unless it's really significant and they have to pay a lot for it,
then they call it "rapacious monopoly").  Their loss.

    Ian> It was obvious, in the sense I meant in the sentence which
    Ian> started this whole particular thread (``most good ideas in
    Ian> software are obvious'').

Oh, it was, was it?  Then howcum they didn't write it that way in the
first place?  They could have, no?  But they had better things to do
at the time.  You picked a right time for a right product.

So tell me.  Today people emulate your configuration system (and try
to beat it, and by now often do, I suppose).  Who were you emulating?
Or did you have to reorganize the configs and documentation yourself?

    >> If the Mac and MS Windows didn't exist, would KDE and GNOME be
    >> anywhere near as developed as they are, if they even existed as
    >> coherent projects?  I doubt it very much.

    Ian> Who knows?  Isn't this another example of ``free software
    Ian> can't do X,'' where X goes through ``good compiler,'' ``make
    Ian> a profit,'' ``make a complete OS,'' etc.?

Depends.  If "can't" implies "never", that's not what I mean.

I'm definitely not saying "never".  I believe there will come a day,
not so long from now, when free software is better than Word (at the
same time), provides all the integration and functionality of Office
(at the same time), etc, etc.  This is true today of Linux/*BSD
vs. proprietary OS kernels for most users, right?  So why doesn't
Linux take over the desktop?  Because it doesn't have a competitive
one yet.  Someday it will, at the standards of the day.

But Microsoft, which probably will exist and be dominant or at least
one of the leaders at that point, won't be worrying about desktop
technology any more, if they do now.  (Except to keep up, so that they
can maintain the seamless integration that is essential to their
monopolization strategy to date.)

What free software (at least FSBs) does not have the capability to do
_routinely_ and _across all software markets_ right now is compete on
the cutting edge of technology.  It can track the technology and force
costs down everywhere.  It can exercise leadership in _some_ markets.
But it can't do both everywhere.

Tracking and competing on price or services is a good strategy,
especially for small consultant-type businesses.  It even seems to
work on a medium scale in some cases such as the commercial Linux
distributors.  The leadership strategy is far more risky (whatever
licensing mode you choose), and thus is hard to sustain without
proprietary revenues.  (This is absolutely true, thus the
Constitutional provision for IP.)

    Ian> You may be right, but, in fact, we don't know.

Of course not.

The point is, when somebody hands you a million dollars and says, "go
build a business", what business are you going to bet on?  I'm saying
that some of these things look like rather bad bets to me.

Institute of Policy and Planning Sciences
University of Tsukuba                    Tennodai 1-1-1 Tsukuba 305-8573 JAPAN
              Don't ask how you can "do" free software business;
              ask what your business can "do for" free software.