Subject: Re: 100 Acre Software...a look back
From: Ian Lance Taylor <>
Date: 26 Jun 1999 21:47:08 -0400

   Date: Sat, 26 Jun 1999 11:56:44 -0700
   From: Michael Tiemann <>

   Which leads me to my second observation.  When Yggdrasyl (sp?), Red Hat,
   and other Linux distributors hit the scene, we completely dismissed
   them.  Sure, Linux was open source, and yes, if the AT&T vs. BSD thing
   didn't end soon, there might be a market for Linux.  But these
   distributors violated the lesson we learned in the 100 Acres war.  They
   were doing packaging, not development.  We didn't see the long-term
   position in that.  Did we learn the wrong lesson from 100 Acre?  Or will
   Red Hat have to completely invert their business model (from
   distribution to development) in order to survive?  Or will Red Hat (and
   the other Linux distributors) escape death because no company can create
   a single definitive _commercial_ release?  History, of course, will
   answer this for us.  My prediction is that the answer will depend not so
   much on the finest distinction of the definition of "free" nor the
   subtlest interpretation of the GPL, but rather on the fundamental
   question of where the value is, and who is best able to capture it
   before they are run out by a more competitive company or model.

Remember that when Yggdrasil, Red Hat, etc., showed up, Cygnus was no
longer a nimble startup, but had become a large slow moving behemoth.
Well, perhaps I exaggerate, but you see my point.

Also, Cygnus was in a business model which expected large revenues per
customer, and as we all know it's hard to make the switch to low
revenues.  Only recently has Cygnus started selling a low revenue
boxed product.

I believe Red Hat had an important insight, which Hundred Acres, RTR,
and, yes, Cygnus, didn't have: sell the brand.  Red Hat doesn't sell
Linux.  They sell Red Hat Linux.  They have caps, T-shirts, tours,
mailing lists, announcements, etc.  They ought to have a newsletter
and a fan club, and for all I know they do.

In any case, as Russ points out that Red Hat is hiring developers, and
moreover they had to do development from the start, such as the RPM
mechanism.  They also have always had to do development related work,
namely getting all the packages together and making them work
together.  This is nontrivial--think about what a pain it would be to
do it yourself--and as far as I know every distribution does it
differently.  So they do add real value--not as much value, in my
opinion, as Cygnus adds, but that is reflected in the order of
magnitude difference in price.

I don't have any particular conclusion here, except that I haven't
seen Red Hat make a mistake yet.

(I did buy a Red Hat distribution, and I actually sent in a support
question via e-mail, which was never answered.  That was sort of
annoying, but even that was probably not a mistake on their part.
Offering support is an important selling point, but there's not much
benefit to actually providing it, since that would be expensive.)