Subject: Re: the walls have ears
From: "Tim O'Reilly" <>
Date: Sun, 30 May 1999 18:27:56 -0700

"R. Brock Lynn" wrote:

> But I don't have access to the Solaris source code... It forked to no abandon.
> Perhaps Sun will scrap their proprietary model for the pure sake of the further
> development of the code, otherwise I see SPARC and UltraSPARC GNU/Linux or
> GNU/HURD taking its place in the not so distant future. I've run SPARC GNU/Linux
> before, and liked it so much, I scrapped Solaris altogether. That machine is
> still today, after more than 2 years, the main DNS server among other things for
> a previous ISP company I used to work for.

I agree that this is a drawback of the BSD license.  But being
realistic, how big a drawback is it?  I like FreeBSD better than
Solaris anyway!  And I like a lot the fact that Sun (and other
UNIX licensees) was able to build a large, scalable business that
has had all kinds of spinoff benefits for users.

Heck, the reason I get somewhat irked by "true religion" about
licenses is that we have all received so much damn benefit from
commercial activity under *every* kind of license (including
proprietary licenses) that I think it's somewhat disingenuous to
argue exclusively for completely free software!

I don't think any of us (and particularly a lot of people in the
free software environment) understand the complex business
ecology we live in well enough to pontificate about it as much as
we do.  After all, if all software should be free, this means
that Intel firmware should also be free--and if it were, it's
certainly possible that Intel would still have flourished, but
it's also possible that they wouldn't have had the ability to
make the enormous investments in plant that are required to make
a new generation of chips.

Just a for instance.  It's easy to play guessing games about the
past, but it's better to do so with some facts, and this whole
discussion is long on opinion and short on facts.

What we know is what *has* happened, and we can analyze *why*
it's happened.  Analyzing what *might have* happened is a lot
harder to bring to a conclusion.

I believe that there have been enormous benefits (and major
drawbacks) from trade secrets, copyright and other forms of
restricted intellectual property.  I also believe that there have
been enormous benefits (and some drawbacks) from each OSS license
out there.  

Our goal should be to find what works by experiment, and by the
study of history.

Right now, some companies (Red Hat, VA, Cygnus etc.) are out
there inventing the history that we should be studying if we want
to understand the impact of the GPL.  It hasn't happened yet on a
big scale, though.  

The history of what has been sparked by work under the BSD
license (particularly the internet) is more visible.

But I digress.  I guess I could say "the only thing we learn from
history is that people will always argue about history."
> To the average Joe consumer, Sun is just like Microsoft. In fact you could argue
> that Microsoft has benefited from "free software" due to the many UNIX design
> features that are a part of "DOS" and "Windows". That is if you consider the
> early UNIX from AT&T to be "Free Software". UNIX was the first big jolt in the
> general direction of "free software" wasn't it? Of course it was partly due to
> the Anti-Trust restriction on AT&T from entering a software market.

Absolutely.  I do argue that MS has benefitted enormously!  As
have companies from the Internet economy (from Amazon to Yahoo!),
most of whom haven't given huge amounts back to the free software
community either.

> Yes, that's why I tend to dislike proprietary code-forks. 1) It fragments the
> market, and 2) Takes away freedom with respect to the source code.

I think it's somewhat unproven that it fragments the market.  And
it's also unclear that forking is a bad thing.  Apache was a fork
from the NCSA server.  That was good, not bad.

Nor do I agree that it takes away freedom.  It simply doesn't
give back to the soil.

I tend to think of the analogy with the environment a lot.  GPL
licenses are like organic farming.  They give a lot back to the
soil.  Proprietary licenses are like pure chemical farming. 
Berkeley licenses are like plowing a lot of great compost into
the soil, but not being religious about what people grow or how
they grow it after you do that.

I happen to like organic.  But in my own garden, I don't hesitate
to spray from time to time if I have to.
> You lose the free bug fixes (security as well as technical) and enhancements
> that could potentially be sent in by the users. It becomes very easy for the
> non-free code to contain back doors, and unauthorized surveillance. It makes it
> easy for the company to become involved in deceit. And because the code is
> unavailable for close scrutiny, it is more likely that crackers will find
> security holes, from intense reverse engineering, before the company or the
> users do, to their detriment. We need to keep those "2600" readers at bay... ;^)
> without, of course, restricting their rights to speak about and read about
> security holes, and methods of exploit.

All these things assume that you have to use the proprietary
code.  If it isn't better, don't use it.  If it has these
problems, don't use it.  No one is forcing you to.

> Fellas, there's gold in them thar free software hills. And the hill with the
> most of it, that can provide the most in the long term is the hill marked with
> the big GPL. :)
> But of course don't take my word for it, study the situation carefully, and draw
> your own conclusions.

It is certainly true that there is enormous gold in the GPL (and
Linux specifically) right now.  When Red Hat and VA go public,
this whole area is going to explode.  But this doesn't mean that
this is the only or the best way to go.

After all, Dell did a better job than Compaq of exploiting the
pure dynamics of the open PC hardware platform, but this doesn't
mean that Dell was a more successful company than Compaq
(although it is doing better at the moment.)

My point is that there's gold in them thar free software hills,
and anyone who is only panning the GPL stream (OR the Berkeley
stream), or for that matter only the software stream, is as
likely to strike out as to hit the big  time.  After all, the
most enduring business left over from the California Gold Rush
was the Levi-Strauss company--not a gold mining company at all.
> . Of course short term oddity companies like Microsoft throw off
> the gullible business men who think that Microsoft has perfected the process for
> long term economic growth. When in fact (or rather at least I don't believe so)
> it has not, but rather has acquired its wealth "on the margin" of an infantile
> new market... That of software production. The market has yet to "stabilize" so
> to speak, and is under considerable "growing pains". Perhaps we are all just
> beginning to feel the "labor pains", and the true "baby" hasn't yet been born.

That I agree with completely.  But most fortunes are made at
these kind of transition periods.  I don't think it's fair to
beat on Microsoft.  They did us all a lot of good, believe it or
not.  They just outlived their usefulness, and have been trying
to hold back the next paradigm shift.

> True. But this makes it easier for new businesses to enter the market place, and
> spurs competition.

But too much competition can be as bad as too little, especially
in an industry that requires heavy investment.  And believe it or
not, there are some areas of the industry that do require it.

> Perhaps for products made from scarce resources. But what about information, or
> knowledge based products? (that once created are easily and cheaply duplicated)

I give you Intel:  a knowledge-based company that uses commodity
materials, and yet requires ENORMOUS investment to keep on

Tim O'Reilly @ O'Reilly & Associates, Inc.  
101 Morris Street, Sebastopol, CA 95472   
707-829-0515 ext 266, Fax 707-829-0104,