Microsoft's Steve Ballmer said yesterday (7 April 1999) that Microsoft
might consider offering some of its Windows code as open source if the
Linux operating system continues to increase in popularity. We in the
open-source community welcome this development. We are confident that
Linux will continue to attract users who value reliability and
flexibility, and Mr. Ballmer's undertaking may point the way to a future
in which the powerful quality and reliability benefits of open source
are enjoyed by users of *both* of the two most widely used operating
systems in the world.
We'd like to remind Microsoft that (as Jamie Zawinski put it recently
in his Mozilla resignation announcement) open source is not magic
pixie dust. Code that's badly designed or non-functional won't
instantly improve simply by being open-sourced. Before the peer-review
effect can benefit consumers, lots of developers must be both able and
motivated to participate. We must therefore caution Mr. Ballmer and
Microsoft that empty demonstrations and half-measures won't do.
A partial release of components that won't build into functioning,
usable software won't attract developers. A release of "Windows" that
leaves the kernel, the Windows API or critical pieces such as Active
Directory, SMB, OLE/DCOM, or the Exchange wire protocol still closed
will readily be diagnosed by both developers and the Justice
department as a sham. So would a license that exposes source but
denies outside developers full rights to modify, re-use and
re-distribute without legal hindrance.
These are all traps to be avoided. But if Microsoft is sincere in
wishing
to join the open-source community, and does the right things in the
right
spirit, we will welcome it. Truly open-sourced Windows code would be a
boon to consumers and developers everywhere.
Signed,
Eric S. Raymond, President, Open Source Initiative
Dr. Larry M. Augustin, President and CEO, VA Research Linux Systems
Russell Nelson, President of Crynwr Software, OSI Board member
L. Peter Deutsch, OSI Board member
Larry Wall, inventor of Perl
Guido Van Rossum, inventor of Python
-- Real programmers don't document. If it was hard to write, it should be hard to understand.