Subject: Re: "incentive void" (was Re: A different patent covenant...)
From: "Stephen J. Turnbull" <>
Date: Wed, 06 Dec 2006 01:06:37 +0900

Thomas Lord writes:

 > I've sat in meetings between such companies and the purchasing
 > reps of such customers.   These meetings, when they are between
 > inexperienced start-up reps and experienced customer reps, are
 > downright funny.   I mean belly-laugh funny:

Well put analysis.

 > All of this is to say that, because IP isn't strongly central to making
 > sales in this area, but quality-of-team is central:

Correct.  The real issue is whether IP has a role in keeping the team

 > So, why doesn't it happen?   Why aren't there a bunch of pure, virtuous,
 > free software businesses in the appliance industry?  Well:
 > You, Stephen, suggest it doesn't happen because of the lack of IP
 > incentives.

I can see how you'd read that from what I've written, but let me
defend myself by saying I haven't done my own opinions justice. :-)

Specifically, if you review everything I've written, you will find
that I have argued that IP is on the one hand what developers can sell
to managers, and on the other, is what managers can use to control
developers.  Since actually asking you (or me!) to review that huge
glob of verbiage is absurd, just take my word for it, it's there. ;-)

 > I disagree.  Raymond popularized, and made central, the "magic
 > cauldron" view of open source.  Later, Greg Stein and others
 > elaborated this conception.

I don't know Greg Stein's claims.  I do know Raymond's.  Raymond's
essay on the magic cauldron in many ways anticipates the themes you
present in your post.  I'm not sure why you think he thinks it applies
to network appliances; I think he explicitly sets conditions that
exclude them.

 > In the magic cauldron view, you Stephen, should take your vague
 > idea for a network appliance and "release early, release often".  

I don't think that Eric would agree that network appliances would work
that way.  In his Ch. 9 taxonomy, they'd come closest to the "widget
frosting" model---but in the appliances you describe the software is
not frosting, it's the force that through the circuit drives the
current.  This violates the basic condition for the widget frosting
model: "The software itself is not a profit center.  It's an overhead
-- often a substantial one."

Specifically, I think the network appliances you've described violate
all but condition (e) of Magic Cauldron, Ch. 10.2, and violate (b),
(c), and (d) quite strongly.

 > For the purposes of your analysis of the appliance biz, Stephen,
 > the magic cauldron is a failed concept because, while it might
 > succeed in producing a release of useful software, it is outright
 > hostile to the formation of that "team of 10-15" a high-end
 > customer is looking to buy from.

I could have passed the quiz if that had been the only question on it,
Tom!  Honest, I already knew that.  And so does Eric, I imagine.

 > but the cauldron modes of organization fail to present that team
 > to customers as a legal entity with whom to do business.

That's right.  The magic cauldron is not about FSB as legal entity.
It's about legal entities that would rather be playing golf than
writing software.

 > Yes, absolutely, *share* your software according to the golden
 > rule.  Yes, absolutely, use a copyleft license to preserve freedom.
 > Just leave the damn volunteers out of it if your aims are for
 > private gain.  Don't count on the "magic cauldron" to do anything
 > other than (a) contribute weakly to publicity; (b) otherwise be a
 > nuisance.  Such an assumption will leave you never disappointed
 > and, occasionally, happily surprised.

Nice turn of phrase.

But "just leave the damn volunteers out of it" begs two important
questions.  1. How do you keep the team together if the ones who think
they're doing all the work decide to walk?  2. How do you placate the
infantile VCs who keep crying until you stick "patentifiers" in their
mouths?  I think that IP has a role to play in both, though at this
point I don't have a clear conception.  (Admittedly, if we abolished
IP, the second role would be moot.)