Notes from Eric S. Raymond’s Presentation to the
Philadelphia Linux Users Group
Eric gave the group a quick summary of his previous two papers and a preview
of his third paper "The Magic Cauldron." The theme of his third paper
is based on the sociologic and economic motivations that were the impetus for
the creation and development of Linux into its current position as the single
most technologically sophisticated and reliable operating system ever
created.
Eric began by comparing the 3 methods of coming into ownership of property
under Anglo-Saxon law with the 3 methods of acquiring ownership of software. He
drew analogies between pioneers homesteading a previously uninhabited/unowned
property to founding a new software project. He equated the transfer of
ownership of a piece of property to the previous owner of an existing software
project transferring (selling or giving) that software to a new owner. And
finally, he compared the user who comes to own software that is not currently
owned or maintained by anyone to the common law concept of acquiring ownership
of land through adverse possession.
Eric seemed to be surprised by the parallels. After all, he pointed out, what
could be more dissimilar than tangible, finite property with intangible,
infinite software code.
From a sociologically point of view, he opined that the creation of the
concepts of territory and property are conflict avoidance devices. He argued
that anthropologists who state that not all cultures have property rights are in
error. Even the bushmen who have no defined hunting territories, do own and
defend their water holes. Value of an item is relative to its scarcity and the
ability to defend the item. Further, the better the quality of real property in
terms of its ability to yield product, the more advanced the property rights
system, across all cultures.
Eric thought that the concept of "forking" (where versions of a
software project diverge into their own entities) is a taboo within the hacker
culture. He thought that the time and energy people spend defending their
actions after a software fork belie the fact that the culpable parties know they
are violating a social norm. This concept was met with considerable scepticism
by the audience.
- If cultures are adaptive and people will defend their title to property
because of the yields of ownership, then how does one explain the emergence
of the open source model of software ?
- Peer admiration.
There are 3 types of cultures, and within each, man (and woman - don't give
me a hard time here. "man" means both man and woman) competes for
social status due to existing evolutionary tendencies. In the command hierarchy,
he with the biggest club or control of those with the biggest clubs controls
others and has the highest social status. However, the command hierarchy
doesn’t scale very well. As this type of community grows, the lower ranked
individuals tend to create inefficiencies through their designs on higher
ranks.
In the exchange economy, he with the most to trade has the highest status.
But note that the command hierarchy can exist parasitically on top of the
exchange hierarchy. Last is the gift culture in which he who gives away the most
gains the greatest stature. Such cultures existed in the Pacific Northwest and
tend to only thrive in environments in which the members are
"wealthy," i.e. they have more than their basic needs for survival and
exist in a relatively benign environment. The impetus for this behavior,
according to Eric, is also evolutionarily disposed. Eric’s
anarchic/libertarian views were evident in this portion of his speech.
(One wonders how a gift culture can exist for an extended period of time.
Those ecosystems that would foster the gift culture, i.e. the richest
communities are the most subject to a phenomenon that tends to equalize relative
wealth among cultures of various levels of wealth: reproduction. In those
natural systems which have the most resources available to its population,
reproduction, followed by rapid population growth tends to lower the wealth of
the average member to a level similar to that in ecosystems of average or even
below average resources. In fact, such rich environments often lead to
population explosions creating an ecosystem rather low in average wealth, which
leads to a population implosion).
The Hacker Culture is quite similar to the gift culture, with members
competing for status through their contributions to the source code. It is an
"epi-phenomonon" which floats on top of an exchange culture because
the exchange game does not motivate Hackers. And the hackers have both the time
and knowledge to compete with each other. However, the Hacker Culture is a more
pure type of gift culture, because the "gifts" have absolutely no
exchange value whatsoever.
Economics of Open Source
Why doesn’t Harden’s Tragedy of the Commons apply to Open Source
? Or, in other words, if an asset is "free" isn’t it always
misused or aren’t resources misallocated? Eric argues "No." The
difference is that Open Source contributions actually improve the Commons,
whereas in Harden’s model, the Commons are rapidly depleted through
overuse. (What Eric misses here is that the hackers are not necessarily the same
as the cattle owners who over-graze the commons because the hackers are not
necessarily the end users). Thus in Open Source, there is an inverse commons
effect. OS is a positive sum game, whereas the commons is negative sum.
Eric also discussed an economics game in which a professor puts $5 into a jar
and doubles the amount still in the jar every 5 minutes. The students are free
to take the money out of the jar at any time. In theory, the participants act
selfishly the first few times the experiment is run, but then learn to cooperate
to maximize their individual self-interests. (Eric failed to point out why those
in Harden’s Commons are unable to come to a similar cooperative
agreement). If you can cite the source for this economics game, Eric wants
to hear from you, since he can't find it.
The inverse commons model and the economics game will form 2 of the 3 themes
of "The Magic Cauldron" with an addition in which Eric discusses the
circumstances under which a negative sum game can turn into a neutral or
positive sum game.
He then discussed Open Source advocacy within commercial enterprises.
Comments from moi (jas3):
- Eric has correctly identified the motivations for the Linux
community’s contributions to the source code. Peer admiration and the
ability to contribute to a body of existing code are powerful motivations
for the hackers and wizards responsible for its creation and development.
Once it gained a critical mass, contributing to the creation of Linux within
the Hacker culture became an honor for the most technologically savvy
coders. Plus, there is a certain aesthetic to working on
something which is technically beautiful.
- Eric did not discuss the needs that Linux was able to fill. It is an
educational device for those studying Operating Systems. It is a tool for
those who cannot afford the licensing fees Microsoft requires.
- However, he misses point that many of the programmers have other
incentives. Educators have created and use Linux to teach their students how
operating systems function. Much functionality has been added, not by those
who are trying to impress others, but by those who need that functionality.
This need often arises for the hobbyist, such as the extensive support for
Ham radio stations, but also arises within the business world.
- The comparison of the Tragedy of the Commons to the development of Linux
misses a point. The cattle owners who are grazing the Commons are competing
for a scarce resource. The programmers who created Linux have an abundance
of resources that allow them the time to compete with each other for status
within their programmer community. The grazers are fighting for their
survival. The coders are writing in their spare time (when not playing
Doom).
- The development of LINUX for personal computers and servers has been
largely possible through the incompetence of the commercial software
developers, chiefly Microsoft. The application developers are hardly better
than Microsoft, which has to its advantage a de facto monopoly not available
to most application developers.
- The PLUG’s and Eric’s hope that Open Source becomes a model
for the entire software industry borders on the religious. Open Source
should and will succeed only if it is a better model for developers and
consumers.
In the end, the commercial software industry could remedy most of the
problems that have led to the emergence of Linux and the OS model if they would
make a few changes. Although these changes will be hard for the software
publishers to adopt, those that do will survive. Those that don’t,
…
Most application developers would be wise to adopt some of the more
successful practices of Linux and the Open Source model, such as:
- The rapid identification and correction of bugs in software, followed by
the immediate distribution to existing clients of the corrected software
at no additional cost.
- The release of the source code to the customer, with licensing rights
that allow customer to make changes that suit his business needs. Or at
least, the release of the source code to a select group of customers who
have the desire and the technical ability to review it. Perhaps these
customers could be the same customers who Beta test commercial software.
Perhaps the software publishers could incentivize the customer base by
rewarding those who found (and corrected) bugs economically (free software -
trips to EuroDisney - cold hard cash).
- The creation of user forums to which employees of the publisher can and
do contribute the solutions to problems.
- The release of new software at no (or at least minimal) cost to new
users, e.g. Vueprint at Hamrick Software.
- To the extent that today’s commercial software developers can
become more responsive to their customers needs more rapidly, Open Source
software may be a solution looking for a problem. I am willing to concede
that Linux may be an exception to the rule, but only because Microsoft has a
monopoly in operating systems. This monopoly gives MS the ability to
distribute buggy, bulky, slow, poorly documented, overpriced software, for
which there is no competition. MS has been able to leverage its OS monopoly
into creating a host of application programs that through predatory pricing,
coercion, theft, and purchase of its competitors have driven much of its
competition out of business. It’s practices in polluting the JAVA
language were driven for one reason only: to kill the thin client
applications that would (or might) render its OS superfluous.
I recognize that the last bullet is anathema to most PLUG members and to the
general OS community. And I wish you luck, but in the business world, I
believe that the commercial application developers WILL wise up. Word
Perfect, for instance, did release a Linux version, but I don't think they'll
ever release the source...
Comments, Insults, Flames welcome to either here (assuming the moderator
approves) or to me at:
jas3@netreach.net