Subject: two FSB growth opportunities
From: Tom Lord <lord@regexps.com>
Date: Thu, 4 Oct 2001 15:32:19 -0700 (PDT)



Large FSBs are building web sites -- initially just for providing
upgrades by subscription.  Later, I'd predict, those sites will become
far more content-rich, and will include forums linking the customer
community to one another and to developers.

Using more or less just the FS technology basis set we already have:
If the content of those sites effectively targets individual
engineers, even at a very low price, FSBs can collect larger revenues
at the same or lower cost (compared to consulting and custom
engineering) while providing better service and performing a greater
social good (shall I post a business plan for this? :-).  That's a
path to the next robust level of growth and profitability for the
large FSBs and it largely does away with depending on projects to set
aside budget to outsource support and custom development.  It has the
potential to create a useful channel for small FSBs developing more
specialized components but having trouble finding sales.

Using new free software technology: If the content of those sites
targets broad consumer markets, those sites will generate windfall
profits, even at very reasonable prices.  This is the only route I'm
aware of by which FSBs can take on proprietary and mixed-mode vendors,
instead of getting squashed by them.

I don't think this is overstated: Every technical decision made by
today's FSBs, every dollar invested in R&D, every resource available
to an FSB not consumed by the demands of immediate survival should be
carefully and methodically spent in persuit of creating content-rich,
consumer-oriented, software delivery web sites.

It's not necessarily a problem to try to attack directly, though.
There are technological prerequisites and there is a need to
experiment with smaller scale web sites for delivering software and
related content via the web.

Web sites targetting engineers are a good way to work out the business
models and engineering practices needed for consumer-oriented software
web sites.  What about the technology for consumer markets?

The open source community is doing a so-so job of building consumer
oriented software.  They have quality problems with what they've
built, and architectural problems limiting the future of their
technology.  These are good candidate areas for FSBs to direct
resources.

What kind of architecture would be better than the ones we have?  It
would have to:

	- Be smaller and simpler -- since the resources to build,
	  test, and document it are limited, and only loosely
	  coordinated.

	- Be robust -- "consumers don't grok fsck and spaz trying to
	  clean-up after core dumps"

	- Be highly adaptive -- so that the customer forums on these
	  sites can drive new applications, rather than always
	  having new applications driving customers.

Somewhere in the union of SCSH, the Hackerlab projects, the Emacses,
SCWM, and the original goals of the Guile project one can see a set of
ideas developed by a loosely coordinated community of hackers, over a
period of decades, that fits the bill.  (The "universal web browser"
approach parrots these ideas, but in a form plagued by an avalanche of
complex standards and bloated technology, with a nearly-complete
abandonment of any simple-yet-powerful approach to UI-client
extensibility.  Worse may be better but awful is just plain awful.)
There's no more technical obstacle to building a clean, effective,
commercial quality realization of these ideas: it's something a few,
not-very-large, well organized, and highly focussed teams can do
within a few years of starting, most likely producing useful ancillary
results along the way.

One approach to instituting such teams to is to create independent FS
R&D labs who have large FSBs and mixed-mode companies as customers.  A
simpler approach might be to create such labs within one or more large
FSBs.  

love and hope,
-t