Subject: Re: open source winter
From: Ian Lance Taylor <ian@airs.com>
Date: 2 Apr 1999 14:53:17 -0500

   Date: Fri, 02 Apr 1999 13:59:31 -0500
   From: hecker@netscape.com (Frank Hecker)

   To go back to Kragen's point, yes this is an issue of performance
   against expectations not of performance against reasonable goals.  The
   more general issue here is the critical importance of managing
   expectations, which is something not unique to JWZ and the Mozilla
   project.  To take but the most prominent example, there's a positive
   feedback loop going on right now between the press, open-source
   evangelists, and "anybody but Microsoft" companies regarding the chances
   for Linux world domination in the near term.  If the millenium comes and
   (say) Linux hasn't yet penetrated the desktop in any major way, will
   that be perceived as failure too?

No, because at least in what I read, people consistently say that
there are no end-user applications for Linux, and therefore it's not
clear whether it will penetrate the desktop.

However, if the Linux share of the server market slips, then it will
be perceived as a failure--a good try, but Microsoft wins anyhow.

And if large companies back-pedal on releasing their sources as open
source, then open source will be perceived as a failure, even though
the same GPL/BSD free software will still be there, working just as
well as it ever did.  That's the scenario I expect to happen: the hype
curve has built up too high, any high level of hype invites a
backlash, and the backlash always pushes things down too low.

I think the effect will be to return free software to roughly the
state it was in five years ago: perceived as hackerware or hobbyist
code, not suitable for use in serious commercial enterprises.  That's
the version of open source winter that I see.

If I were a free software spokesman to the press, I would be working
to lower expectations, to stick to the proven story--existing free
software is highly reliable--and avoid the unproven ones--that free
software can be end-user friendly, or that adequate end-user support
is available, or that businesses can do well by freeing their source
code.

Ian