Subject: Re: [may be junkmail -pobox] Free Software Bazaar
From: (Frank Hecker)
Date: Mon, 28 Sep 1998 11:18:25 -0400

Kragen wrote:
> On Fri, 25 Sep 1998, David Welton wrote:
> > I like the idea, but I wonder if one wouldn't leave feeling a bit
> > cheated, having paid for something that everyone subsequently gets for
> > free.
> Hmm.  It seems that, if this is a problem, it is a problem that will
> arise with most FSB business models.

And not just FSB business models either.  You'll find this situation in
any environment where product prices are being driven down by either
technical innovation or competition or both.  Compare those who bought
Pentium PCs when they were USD 5,000 instead of USD 1,000, or for that
matter those who bought web server software early on.

The buyer who pays more in the early days (before the price goes down)
presumably does so because they derive value from the product, and can
justify paying the price for that value at the time they buy it.  If
delay buying to take advantage of expected lower prices then they pay
the opportunity cost of not benefiting from the product's value during
the delay.  If the product is such that it can save the customer a
considerable amount of money (or make them an equivalent amount of
money) then this opportunity cost can be quite large and thus the
benefits of buying early quite substantial.

> This is discussed in the excellent Street Performer Protocol paper
> someone posted here about in the last couple of days.  I read it; I encourage everyone
else to read it, too.

That was me who posted the URL for the paper; and now that you've read
it, let me give you one of my thoughts on it:

One of the main ideas of the paper is using a strategy of delaying
release of a product until a predetermined amount of money has been
contributed to the creator.  One of the main areas that the paper
proposes the use of cryptography is in guaranteeing to customers that
there is in fact a product in the first place, and that the product as
released is the same product they paid for.  (This is done basically by
taking a hash of the product's bits and releasing that at the time
contributions are solicited.)

This would seem a reasonable thing to do for works like books and
movies, which are typically released for the first time as fully
finished products.  However it just seems strange to delay release of
software in this way, because you lose a period of time during which
other people could be helping the creator improve that software (e.g.,
by finding bugs, helping to fix them, and so on).  It would seem to make
more sense to release software as soon as it embodies a "plausible
promise" (to use Eric Raymond's phrase), in which case a lot of the
cryptographic protocol in the paper seems like overkill; the proposal
then reduces to the idea (certainly not new to the paper) of donating
money to fund software development, and having that money held by a
neutral third party until some agreed-upon milestone has been met.

Frank Hecker          Pre-sales support, Netscape government sales