The Behavior of Open-Source
Support Companies
Bruce Momjian
DRAFT
Open-source support companies face unique challenges. This article explores
a few of them:
- The unusual inter-company issues faced by open-source support companies.
- The paradox of why, sometimes, doing the best for your company may be the wrong
thing to do.
- Why companies should work together to preserve their shared open-source asset.
Traditional companies have their own marketing, sales, research/development,
and manufacturing departments. They are self-contained organizations that share
very little with other companies in the same market.
Open-source support companies are different. The open-source software they support
is a shared resource. All support companies rely on the health of that
shared resource for their livelihood, and because they rely on it, companies
take actions to maximize the value they derive from that shared resource. However,
these actions can make things worse.
How can acting in your self-interest actually be counter to your self-interest?
The prisoner's dilemma illustrates such a case. Two prisoners are captured
by police and placed in separate cells. The police have enough evidence to convict
each of a minor crime that will result in a one-year prison term. However, the
police know the prisoners have committed a more serious crime. Each prisoner
is told that if he confesses and the other prisoner does not confess, he will
go free, and the other will receive a twenty-year prison term. If they both
confess, they will each receive a ten years prison term.
For each prisoner, the decision in their self-interest is to confess.
Each prisoner does not know what the other will do. However, confessing produces
better results no matter what the other prisoner does:
- ``Suppose the other prisoner confesses. If I confess too, I get ten years instead
of twenty.''
- ``Suppose the other prisoner does not confess. If I confess, I go free
instead of serving one year in jail.''
The interesting effect of the prisoner's dilemma is that each prisoner, acting
in their own self-interest, produces a worse result, two ten-year jail terms,
than if both had not confessed and gotten only one-year jail terms.
The prisoner's dilemma, first formulated by Albert W. Tucker in the 1950s, has
been applied to many fields, including economics, foreign policy. and philosophy[Blumen].
The prisoner's dilemma even applies to open-source support companies. Each company
is like a prisoner in a cell. Each wants to dominate the open-source community,
and fears other companies will do the same. Unfortunately, domination by multiple
companies only diminishes the health of the open-source community, yielding
a worse result than if they had not acted.
Perhaps dominate is too strong a word, but companies do position themselves
to receive maximum benefit. When all companies do that, they can destroy the
shared resource they rely upon. In prisoner's dilemma terms, they receive ten
years in jail instead of one. They reason, ``If I dominate the shared resource,
and the other companies don't, I win. If they do, and I don't, my business suffers.''
Unfortunately, if they both do, the community suffers, and the companies along
with them.
The good news that the prisoner's dilemma is not played just once. It is played
by open-source companies over and over again, in the little and big things they
do that affect their shared resource. And with repetition, there is hope. When
companies realize how their actions to control the shared resource cause other
companies to do the same, an arms race occurs. And once they realize
that, they can start to seek a truce, where companies respect the shared resource,
rather than dominate it at every opportunity. With such cooperation, companies
get the maximum benefit, because the shared resource remains healthy and vibrant,
and all companies prosper.
This paper describes how open-source software companies can analyze their actions
in cases the affect the open-source community. Restraint is often the best practice.
Fortunately, 99% of a company actions have no affect on the open-source community,
so they can behave just like normal companies, seeking to grow and prosper.
The Behavior of Open-Source
Support Companies
Bruce Momjian
DRAFT
Open-source support companies face unique challenges. This article explores
a few of them:
- The unusual inter-company issues faced by open-source support companies.
- The paradox of why, sometimes, doing the best for your company may be the wrong
thing to do.
- Why companies should work together to preserve their shared open-source asset.
Traditional companies have their own marketing, sales, research/development,
and manufacturing departments. They are self-contained organizations that share
very little with other companies in the same market.
Open-source support companies are different. The open-source software they support
is a shared resource. All support companies rely on the health of that
shared resource for their livelihood, and because they rely on it, companies
take actions to maximize the value they derive from that shared resource. However,
these actions can make things worse.
How can acting in your self-interest actually be counter to your self-interest?
The prisoner's dilemma illustrates such a case. Two prisoners are captured
by police and placed in separate cells. The police have enough evidence to convict
each of a minor crime that will result in a one-year prison term. However, the
police know the prisoners have committed a more serious crime. Each prisoner
is told that if he confesses and the other prisoner does not confess, he will
go free, and the other will receive a twenty-year prison term. If they both
confess, they will each receive a ten years prison term.
For each prisoner, the decision in their self-interest is to confess.
Each prisoner does not know what the other will do. However, confessing produces
better results no matter what the other prisoner does:
- ``Suppose the other prisoner confesses. If I confess too, I get ten years instead
of twenty.''
- ``Suppose the other prisoner does not confess. If I confess, I go free
instead of serving one year in jail.''
The interesting effect of the prisoner's dilemma is that each prisoner, acting
in their own self-interest, produces a worse result, two ten-year jail terms,
than if both had not confessed and gotten only one-year jail terms.
The prisoner's dilemma, first formulated by Albert W. Tucker in the 1950s, has
been applied to many fields, including economics, foreign policy. and philosophy[Blumen].
The prisoner's dilemma even applies to open-source support companies. Each company
is like a prisoner in a cell. Each wants to dominate the open-source community,
and fears other companies will do the same. Unfortunately, domination by multiple
companies only diminishes the health of the open-source community, yielding
a worse result than if they had not acted.
Perhaps dominate is too strong a word, but companies do position themselves
to receive maximum benefit. When all companies do that, they can destroy the
shared resource they rely upon. In prisoner's dilemma terms, they receive ten
years in jail instead of one. They reason, ``If I dominate the shared resource,
and the other companies don't, I win. If they do, and I don't, my business suffers.''
Unfortunately, if they both do, the community suffers, and the companies along
with them.
The good news that the prisoner's dilemma is not played just once. It is played
by open-source companies over and over again, in the little and big things they
do that affect their shared resource. And with repetition, there is hope. When
companies realize how their actions to control the shared resource cause other
companies to do the same, an arms race occurs. And once they realize
that, they can start to seek a truce, where companies respect the shared resource,
rather than dominate it at every opportunity. With such cooperation, companies
get the maximum benefit, because the shared resource remains healthy and vibrant,
and all companies prosper.
This paper describes how open-source software companies can analyze their actions
in cases the affect the open-source community. Restraint is often the best practice.
Fortunately, 99% of a company actions have no affect on the open-source community,
so they can behave just like normal companies, seeking to grow and prosper.
- Axelrod
- Axelrod, Robert, The Complexity of Cooperation: Agent-Based Models of
Competition and Collaboration, Princeton University Press, 1997.
- Blumen
- Blumen, Jonathan, The Ethical Spectacle, http://www.spectacle.org/995/.
- Britanica
- Encyclopedia Britanica, <#54#>``The prisoners' dilemma,'' http://www.britannica.com/bcom/eb/article/5/0,5716,117275+10+109420,00.html?query=prisoners
- Helfrich
- Helfrich, Serge, The Prisoner's Dilemma, http://www.xs4all.nl/ helfrich/prisoner/.
- Rawles
- Rawls, John, A Theory of Justice, Belknap Press, 1999.