Subject: Re: "Reasonable Profits"
From: Craig Burley <burley@gnu.org>
Date: Tue, 1 Sep 1998 05:15:41 -0400 (EDT)

>Craig Burley
>commented for example that incredible returns from software will only
>be around for "10-20 years", therefore he's not interested.

Whoa, that's not what I *thought* I said at all.

I was *trying* to point out that depending on proprietary distribution
as an easy means to get non-linear ROI is not likely to last for a
long time, since, among other things, it assumes that end users will
be willing to pay for that for a long time.

That's only part of why I'm not interested -- in investing the next
10-20 years of my life doing *proprietary* SW development!!  (The
other parts have to do with not being able to figure out how to
solve some of the problems I'm interested in solving without incurring
huge development costs...while the libre-software approach seems
to bring forward various interesting, feasible, solutions.)

I look at today's software industry a bit like how some people look
at Internet stocks (Yahoo, etc.), just on a larger scale of time.

The question for Internet stocks is, I gather, something along the
lines of "just what happens if people decide they can get at the
things they want without going through those particular portals"?

The question for software stocks I have is, "just what happens if
end users decide they can get better-quality software in the long
run by insisting it be libre?"

Not that investing in Internet or software stocks for the medium
term isn't worthwhile, at least, but, as an extreme example to
illustrate how I think about that latter question, consider what
would happen if the US Congress magically waved a wand declaring
that all software is libre (say, PD).

My guess is that Microsoft stock would sink to near zero within about
two days, while Cygnus would, if anything, appreciate in value,
probably significantly...as would RedHat and others.  That is,
they're not built on the proprietary model as Microsoft (presumably)
is, so they can more rapidly adjust as the market values libre
software more (which I *like* to think is inevitable, but...).

Anyway, my original reason to make my point was that, in discussions
about how to obtain non-linear ROI with libre software, consideration
should be given to just how feasible obtaining non-linear ROI with
*proprietary* software is in the long run.  I don't see a whole lot
of feasability there.  Which, to me, makes libre software a basis for 
worthwhile business models in the long run, because it ends up not
having such a large disadvantage vis-a-vis proprietary software.

As I said, this isn't based on a whole lot of analysis.  For example,
I observe that the present market "attitude" on software ROI is based
on a model that basically didn't exist until around 20 years ago,
in a field that didn't exist until around 50 years ago.  I speculate
that it is highly unlikely that mankind, still tweaking business models
for growing, distributing, and selling *food*, has managed to converge
on the ideal business model for doing the same with software -- and
that, since the current model assumes that good programmers are
scarce and end users don't understand software well, the long-term
direction is towards less tolerance among the end-user community
for paying high prices for each copy of a trivially copyable product.

(I think we're beginning to see evidence for wider-spread consumer
understanding of copyability and other issues in, e.g., the DivX-
vs.-DVD issue.  Put another way: if DivX and DVD had appeared 20
years ago, I suspect there'd have been lots less understanding of
the issues, and thus more [IMO blind] acceptance of DivX.)

But that's just an example of how I've been looking at this whole
issue for awhile.  There's lots more that needs to go into figuring
this out (on my part at least) before any of us jump to conclusions
(like, say, "get out of software", which John seemed to think I was
somehow concluding).

As far as how quickly you and others have made money in software:
that's all well and good.  People have also made quick money in
panning for gold, in digging for oil, and so on, especially at times
when there was lots to be found and few people that were really willing
and able to find it.  Ghost towns were built out of such successes!

Software is infinitely creatable, so it presumably won't experience
any changes quite that stark, but the fact is that making lots of
money in a short time in a new field has, AFAIK, never been a reliable
way of establishing the long-term viability of that field.

As far as how reliable my prognostications are?  Well, I didn't
believe my "elders" when they (well, one) told me going into
programming as a career would be a waste of time because all the
programs needed by the industry would have been written by the
time I graduated college (12 years later or so).  Nor do I believe
them now when they say "non-linear ROI will `forever' be a property
of selling software, so libre software cannot ever gain a real
foothold in attracting investment dollars, since it cannot assure
non-linear ROI".

        tq vm, (burley)