Subject: Re: torvalds
From: Tom Lord <lord@regexps.com>
Date: Mon, 22 Oct 2001 01:06:17 -0700 (PDT)



	Russ Nelson writes:
       > Tom Lord writes:
        >> Am I the only one on the list who is not particularly thrilled by 
	>> actually reading the code?

	> Am I the only one on the list who is not particularly thrilled by
	> discussions which are not germane to the list?  Yes, there are a lot
	> of clueful people on the fsb list (thank you, all of you).  That
	> doesn't mean that they're here because they want to talk about random
	> things.  It means that they're here because they want to talk about
	> free software business plans, models, and schemes.

There really is an FSB relevence.  I'm not just gossiping.

I've been trying to point to a perspective on FSB's here on the list:

There's a public resource: the Free Software and Open Source basis set
of software.

There's a collection of engineering processes that generate that basis
set.

The quality of the basis set is limited by the quality of the
engineering processes.  The quality of the basis set constrains the
business opportunities available to FSBs.

It's a unique strength of open source that the engineering process is
a mostly public process: trans-corporate and exposed.  Unfortunately,
large parts of the existing engineering process are, in the words of
Wavy Gravy, "not specifically good".

Consequently, FSBs are suffering from a lack of technology agility
and, in some areas, quality.  In my reading of where MSFT, IBM, and
even Sun are at, the FSB community is going to get bit in the ass as a
result.

Part of the problem here is the "honesty" or "reality reflection" of
the most popular FSB business models.  Roughly speaking, they are
proprietary software models, modified in the least number of ways to
reflect Free and Open Source licensing practices.  This is a poor
reflection of reality because it is not deeply grounded in the
engineering processes.

There is a "radical" alternative that is more honest and reflective of
reality: that is to separate two concerns.  First, concentrate on
improving the engineering processes (and I'm just brimming with ideas
for that).  Second, study the engineering processes and implement
revenue models around those (and I can go on at length on that, too).

Nearly everything I've said on FSB is a particular instantiation of
those very abstract ideas, including ragging on the code quality of
the Linux kernel.

My pithy note about Linux, though, was sent in the hopes that previous
messages had instilled the abstractions that give it context.  In your
case, I wasn't successful.  Sorry.

Really - I'm grateful for this list, and have no intention of abusing
it.


-t