Subject: Re: the walls have ears
From: John Gilmore <gnu@toad.com>
Date: Mon, 31 May 1999 12:40:09 -0700

Ben Laurie <ben@algroup.co.uk> said;
> I absolutely don't understand (or agree with) this view that publishing
> something under BSD can somehow lead to it being taken away from me.
> How? Show me the mechanism!

Didn't the farcially-named Open Group announce that future X releases
were going to be proprietary?

Only a concerted effort by the free software community, who said we'd
split off XFree86 and continue development from the last free version,
rendering the proprietary version a side-branch, derailed them.  It
was lucky for us that the Open Group is weak-willed; had they been a
major company, we would've had to follow through with a long-term
effort.

Indeed, this is exactly the point of Microsoft's declared "embrace
and extend" strategy.  Start with the free tools and protocols that
everyone uses; add proprietary features to them; and make the world
dependent on the existence of those features.  The BSD license supports
this strategy; the GPL undermines it.

I hear Microsoft is currently trying to do this to BSD-licensed BIND.
The Windows 2000 release will require undocumented, proprietary
features in whatever name server serves the domain containing the
Win2000 box (this is true of recent betas).  Will most organizations
refuse to install Win2000 boxes, or will they replace their
BSD-licensed name servers with proprietary ones?  Or will they do
something else, like reverse-engineer the proprietary features,
write plug-ins for Win2000 that eliminate the dependency, etc?  You
can predict, I can predict, and then the world will show us what
really happens.

I see increasing numbers of such "attacks" on free software and openly
documented protocols in the future.  The answer is simply money.  If
fifty companies supply compatible TCP/IP's, there's little margin for
each of them.  If one company supplied it, and everyone needed it,
they would get much better than average return on their investment.
Microsoft has figured this out, and actively seeks to create new
"required" niches that they own.  See their financials on
edgar.sec.gov; their cost of sales is 8% of their price!  Microsoft
gets a big multiplier on the total revenues they can get for
e.g. Win9X and MS-Word, because they're viewed as "must haves" but
Microsoft owns 'em.  I get about an email every two weeks from some
idiot who thinks everyone has MS-Word, or more likely doesn't even
realize they're sending me email in that format.  I have to send it
back for conversion to an open format from its proprietary ".DOC"
format.  Most people eventually knuckle under and buy the Microsoft
product that reads 'em.  As long as they can evolve these proprietary
standards faster than we (or proprietary vendors) can clone them, they
continue to win huge margins.

You can argue that, OK, since a major effort by free software
activists defeated the Open Group attempt to "take free software
private", and since there are still opportunities to defeat
Microsoft's first initiative in this regard, the system is still
working.  I contend that it's better to use a license which increases
the work needed to replace free software with proprietary software,
instead of requiring large numbers of people to put in large amounts
of effort to periodicaly combat such predictable problems.

	John