From: Ian Grigg <iang@systemics.com>
Date: Tue, 19 Nov 1996 06:16:42 -0800 (PST)
To: cypherpunks@toad.com
Subject: Re: Crypto Bounties
In-Reply-To: <199611190406.XAA26146@alpha.pair.com>
Message-ID: <3291C196.167EB0E7@systemics.com>
MIME-Version: 1.0
Content-Type: text/plain


azur@netcom.com wrote:
>  Why not the free market rate the programmers.  Have a software
>  distribution/payment collection system which requires that the programmer
>  get a completion bond from a 3rd-part insurance company.  If the programmer
>  fails to deliver as promised and/or on-time the donding company pays the
>  'investors' back in full. Since the bonding company's money would be at
>  risk, if the programmer failed to deliver, they have every incentive to
>  conservatively rate the programmers/companies offering ot build SW to spec.

What you are essentially doing is buying reputation I guess, although I
am not familiar with completion bonds, so I the following is mostly
conjecture on their nature.

There are two major feature/bugs with the completion bond.  Firstly, it
requires the bonding company to make a judgement, and this adds a weak
point into the link.  That judgement will then result in a lot of
emphasis on the fine print of the contract, and fundamentally, software
doesn't work very well when fine print gets involved.  The notion of
"completion" especially sits oddly with the scenario described, where
freeware of mostly source form gets distributed to a wide group of
programmers who can adjust minor problems fairly easily.  I would guess
that when EDS delivers a product, there are completion ("penalty")
clauses, but when the Internet was "delivered" there was no such clause.

Secondly, the existance of an intermediary points to a major
inefficiency in the market.  Whilst your joy at paying intermediaries
may vary, historically, intermediaries follow a pattern of building
barriers to entry, raising charges, and slowing innovation.  This is
theory that is generally applicable in practice, and it is for this
reason that the emphasis in my post was on removing the single point of
guaruntee.  In the context, I would also question the nature of a
financial entity that was willing to sell its reputation to
cryptoanarchists.

>  >[slash]  However, I like this viewpoint because it
>  >eliminates the need for judges.  History shows that a good market
>  >microstructure will beat an authority approach in the long run.

By aiming for no single point of judgement, the only solution that we
found was to rely on a multi-shot game, with past reputation leading to
up-front payment leading to confirmation of reputation.  As it happens,
the bonding companies will rely heavily on reputation as well, in its
varying forms, anyway.

snow@smoke.suba.com originally asked:
>  >> Has anything like this been proposed before?
> 
>  Yes, Eric Hughes proposed a broad and structured proposal primarily
>  addressing this manner of market funding for software development (and
>  possibly suitable for other intellectual property creation) at DEFCON IV.
>  My previous comments on completion bonds were taken from his presentation.

Yes, I saw that earlier post, and did a quick search for it but no
luck.  If any one can point us at the paper, that would be useful.

On a slightly related note - is there any interest in setting up a
mailgroup for programmers and designers of markets?  There's plenty of
cash stuff out there but I know of no forums for trading of fungible
items.
-- 
iang
iang@systemics.com