Subject: Re: FSBs' real competitors are consortiums (was Re: ... maintainership)
From: Tom Lord <lord@regexps.com>
Date: Thu, 14 Feb 2002 18:23:31 -0800 (PST)


Wow -- complete and utter disconnect on intended and received
meaning.  180 degrees off.

      I would caution greatly from drawing too many generalizations
      from too small a sample set. Tom's description of Companies A
      and B appear too identifiable, and thus may exhibit too much
      sample bias.

I do my best to mix and match details from more than one company,
attempting to form a generalized picture of the trends -- but yes, the
set of available data points is frustratingly small.


   That said, let's take off the sugar coating and say that Businesses A are 
   "Legimate/Innovative FSBs" and Businesses B are "Parasitic/Marketing 
   pseudo-FSBs".  The FSB model says that you don't have to do *everything* (the 
   way a proprietary software company does), but you've got to do enough that some 
   freely redistributable software (which can be reasonably hacked upon by others) 
   is actually produced.  The Business B depiction suggests that the business may 
   focus not at all on producing any software of extrinsic value, and focuses 
   solely on finding new markets for existing software.

The very opposite: Business B lives in a world where a lot of the
nebbishy details of day to day software engineering are more
entrenched in the public projects, partly as maintainer habits, but
largely as tool improvements.  A business "A" becomes a business "B"
when it is no longer required to make up for those fundamentals -- and
thus is freed to innovate much more aggressively, on higher-level
problems, and as a rather open-ended and premium service to some
potential big customers that I perceive need this service.



   If the above is correct, then this explains why Tom said that Business A has so 
   much money "tied up": it actually has high fixed costs because the software 
   production line requires actual bodies.  Business B gets all that "for free", 
   and is thus highly leveraged.  Business B is free to spend variable costs to 
   develop specific markets.

More specifically, I think a lot of money is tied up because there's a
big energy barrier between the CVS repos and FTP sites for the public
projects and what you can deliver to customers.



   I have made it a policy to put the following chestnut from Gordon Moore (quoted 
   in 1995 Fortune Magazine) in every one of my presentations:

      "Moore's Cannible Principle"

      The whole point of integrated circuits is to absorb the
      functions of what previously were discrete electronic
      components, to incorporate them in a single new chip,
      and then to give them back for free, or at least for a
      lot less money than what they cost as individual parts.
      Thus, semiconductor technology eats everything,
      and people who oppose it get trampled.

   I see free software (particularly GNU software) as being the equivalent 
   principle for software, and the amount of proprietary software that can be 
   consumed is simply enormous.  Thus, I don't see an end to the runway that 
   enables Business A to be successful in the near future.  Thus, I don't think 
   there's a *necessary* force causing a transition to Business B.

I'm not so sure.  I think projects like "carrier class Linux" are a
forbearer of things to come.  If I'm getting the server market right:
the big 3 or 4 have more and more customers demanding linux as an
alternative to proprietary unix.  Additionally, those companies are
increasingly full-spectrum in the sense of serving both big-iron
consumers and small/medium sized businesses.  There's a bunch of other
trends in adjacent markets that I think tie in with that so that it
all adds up to a situation where, perhaps led by those big companies,
we'll see more and more "corporatization" of the processes that
develop the kernel and core utilities.  The big companies are pretty
well practiced at nailing down issues like testing and release
engineering to a point where they are pretty automagic and I think it
won't be long before they bring that experience to the public projects
-- they'll do that in a diplomatic way and the end result will be a
radical lowering of the energy barrier between public projects and
custom distros.  (Larry McVoy's *1993* paper on a "Sourceware Operating
System" is prophetic and I think companies like Sun are just now
catching up to its implications.)


   That said, there are many FSBs doing many things.  Russ Nelson is happy and 
   successful being the best little FSB he can be.  I don't think he'd ever 
   transition from A to B.  Red Hat continues to affirm its open source and GPL 
   principles, and I think we generate enough extrinsically useful software that 
   we're clearly in the Business A space.  

It may be slightly absurd to disagree with *your* take on RH, but I
think you're really in transition.  I think that's where some of your
partnerships ultimately lead, for example.  What I'm trying to suggest
is: don't let the trend get ahead of you -- the big 3 or 4 are now
each playing two strategies.  On the one hand, they certainly aren't
dropping their proprietary unixes or MS.  On the other hand, they're
offering their customers GNU/Linux as a shared-standard alternative
and the news suggests the customers and analysts are just eating it up
(for good reasons!).  A shared-standard platform needs a rock solid
shared engineering infrastructure and that's what making the
transition to business "B" is all about.  (Sustaining and growing
business "B" beyond the transition is a larger story, involving more
companies: I think you really want your brand to be well-positioned
for that.)



   Companies like Caldera, SuSE, and 
   TurboLinux develop proprietary software unapologetically.  

Again, I can't stress enough that business "B" is not about not
developing software.  It's about shifting the focus of development to
areas that are more interesting to more customers.

   In summary, I think Businesses B are not really FSBs at all, any more than a 
   commodities trader in Chicago 

I see where you're coming from - but it's just a side effect of a deep
miscommunication.

-t