🚀 go-pugleaf

tony@ajfcal.cuc.ab.ca (Tony Field) wrote:
> I have no objection to providing "free" operating system software to the
> computing community and have done so on a few occasions - however I would
> be extremely reluctant to even consider making my geological software
> "freely distributable" since my competition would have access to my
> tricks. In many cases, they have more $$$ available and could immediately
> add my software's features to their products. By issuing my sources under
> the GNU license, I would cause the death of my company since I do not
> have the income of many of my competitors - in spite of the fact that I
> might have a better product.

We have found the opposite to be true.  If you put your product under
copyleft, and sent sources to all your customers along with binaries,
your competitors would not touch your sources.  They'd be deathly afraid
that some ``virulent'' copylefted code would get into their product
by accident, and then all *their* sources would have to be freed.  I'm not
joking.  If you sent the sources directly to your competitors, they would
probably be returned unopened.

Instead, your customers would be really happy.  They would fix the
program where it chafes them, improve it where they wish it did some
simple thing, and probably would send you back their changes (*they*
don't want to maintain 'em in each new release).  Indeed, you might
discover some talented people at your customers, who eventually would
consider working for you on the software.  And they'd already be
familiar with it when they arrived.

Some other geological experts somewhere else in the world might decide
that starting with your software was better than writing their own.  So
they would sell it to local companies, and be making enhancements and
porting it to different machines and improving the documentation
and...suddenly you'd be getting improved versions back from the world,
without having to write them yourself.  Now your proprietary
competitors would really be in a panic; your company has effectively
doubled its programming staff, without becoming any more bureaucratic
or spending more money.

Eventually a small division of one of these large competitors would end
up using your software for some project because it was the best for
their purposes.  It would gradually sweep its way through the company,
picking up features as it went.  You would probably even get a lot of
consulting work out of it, since you would know the software best of
all.

This has happened for GCC.  Nobody has stolen optimization techniques
from GCC.  Instead, GCC has quietly been adopted by many users, then by
small companies (NeXT and DG -- well, they're small in the Unix
market), then by tiny companies that do nothing but sell and improve
the software (Cygnus, Hundred Acre), then by divisions of large
companies whose internal software was too cumbersome to modify for an
interesting side project (Intel, Sun, ...).  Intel ported GCC to the
i960; it's now being used for the i860, and who knows what is cooking
in their labs.  Perhaps the premier compiler for the 586 will be gcc,
because their old compiler didn't grok instruction scheduling, and gcc
provides it for free.

We've been living this story for a couple of years now.  There is a lot
to learn and understand in the `free software market'.  Things are not
always as you would predict them to be.  But it seems to be true that
free software will tend to drive out commercial software, where the two
are roughly similar.  Customers simply prefer it.  And as long as you
know a way to make money on the free software (issuing regular updates;
fixing bugs; documenting it; adding specific features for a price;
distributing it; etc) then your company will come along for the ride.
-- 
John Gilmore   {sun,uunet,pyramid}!hoptoad!gnu   gnu@toad.com   gnu@cygnus.com
"It isn't given to us to know those rare moments when people
 are wide open and the lightest touch can wither or heal."