Professor Epstein,
I just finished reading your article in the Financial Times,
linked from Linux Today. I believe you are missing a number of
critical points in your analysis of the viability of Open
Source.
You correctly state programs licensed under the GPL are not put
into the public domain. However your next few statements regarding
the all-encompassing ownership program are misleading. It is true
that many OSS projects are led by committee. It is also true that
contributors are not compensated financially. Additionally, it is
only partially true that users may keep any version of the code for
personal public use. Finally, there is no basic public version of
the program. All versions are public in that any user can make use
of them.
The committee exists to serve its users. The open nature of the
code is an effective check against poor stewardship, much as market
share can be a check against poor development of closed code.
Secondly, users are free to keep, develop and use any version of
the code for any use, not just personal use. This is important
because the permitted development of alternatives helps to greatly
increase the viability of the check, imposed by open code, on the
committee currently acting as steward of the project. Finally, the
code is owned by those that contribute to it. The Free Software
Foundation does request owners of code transfer their copyright to
them to facilitate legal defense of the open code pool, but this is
not a requirement for participation.
There are also errors regarding your analysis of GPL weaknesses.
You state the GPL has not been tested in court. While this is
technically true, the number of successful resolutions to this
problem indicate both a community willing to address the issues and
a solid licensing foundation with enough teeth to make such
resolutions compelling to both parties. It is FUD to say the GPL is
weak because it has not been court tested without also bringing to
light GPL success as well.
The remedy for open code inclusion into closed code is roughly
the same remedy sought for closed code inclusion into other closed
code. Copyright law serves the creators of open code as well as it
does closed code. In this, the remedy is clear; either remove the
offending code, or modify the license of the entire code body, or
seek alternative licensing with the owners of the code. This is
reinforced by the above mentioned community willingness to address
problems.
Your next statement is FUD as well. You incorrectly assume that
part of the purpose of the GPL is to infect closed projects in a
viral way. This is simply not true. Those of us who choose to
license our work under the GPL do not intend to infect closed works
with our own. We want only to insure those who choose to leverage
our work abide by the terms of the license; namely to keep the code
open for all to use. The cost for use of open code is that the code
remain open.
Put simply, the GPL does not allow a free lunch for business
wanting to make use of GPL licensed code. Either they contribute to
the pool in a like manner, avoid use of the code, or seek
alternative compensation and licensing for their project. If these
terms are not acceptable, perhaps BSD licensed code is a better
choice.
Given the above, your suggestion that the Microsoft operating
system would be forced open due to accidental GPL code inclusion is
a fantasy at best, fallacy at worst. Microsoft has the same options
available to it that any other closed source business does; namely
(as stated above),
- open the derivative work with a GPL compatible license,
- remove the offending code,
- seek alternative licensing terms with the owners of the
code.
The same courts that would deny Microsoft exclusive rights to
derivative works based on their code, which would be difficult
given their code is closed, would also correctly enforce the
options detailed above for GPL licensed code.
Note that paying damages for the inclusion is not an option,
unless the owner of the code finds this acceptable.
The assumption that the GPL only applies to those who knowingly
make use of GPL code is laughable. Traditional copyright law does
not grant an ignorance exception! Once anyone becomes aware of GPL
code inclusion in their project, (I assume you are talking about
development use here because your use of the word ‘use’ is not
clear in context.) they have the same options mentioned above to
seek proper resolution to their dilemma.
You state Open Source is akin to a workers commune. This
assumption is flawed because it does not take into proper account
all the compensation both contributors and users receive from the
Open Source process. You write about shares and collective
ownership and monetary compensation as if the Open Source software
itself has no value when it clearly does.
The growing pool of GPL licensed Open Source software represents
significant value for anyone wanting to make use of or contribute
to it. Unlike a work commune producting products, groups of people
contributing to Open Source software pools benefit from the unique
quality of software; namely, next to zero distribution and
duplication costs. These properties allow for the value of the
software pool to exceed the individual value of its
contributions.
For a user, their investment in learning required to make good
use of the software pool yields far greater returns in terms of
software capability. Said user could, for example, perform almost
any computing task they want to accomplish on just about any
capable hardware platform without recurring license fees and EULA
restrictions. Given the low cost of computing hardware available
today, this particular value proposition of GPL licensed code is
quite attractive compared to the high cost of proprietary software
that comes with its own separate investment in education as well.
Further, this user can continue to receive value from their efforts
by passing along their knowledge and methods to others wanting to
begin their own use of the growing Open Source software pool.
For developers, the situation is similar, provided they come to
acceptable licensing terms with the Open code developer community
as detailed above. Developers can leverage the pool of software in
their own projects much as they would other closed software
toolkits.
Companies today are supporting Open Source software development
by funding key developers. They also purchase companies, like
Novell did with SUSE, in order to achieve greater synergy between
their own projects and the open ones. However, none of these
efforts change the GPL nature of the code, while all of them remain
subject to the check posed by the users and other interested
developers.
Because the software remains licensed under the GPL, anyone
moving away from serving the users best interests will be subject
to alternative projects and methods that will restore that balance.
Contrast this with closed source companies who buy competing
products only to bury them in favor of their own at the users great
expense.
All in all, the number and nature of the errors in your analysis
does not lend credence to your conclusion. Open Source software has
and will continue to change the very nature of software
development. It also leads the transition from selling software
solutions to software services. This transition will see the value
of large closed software development houses diminish, unless they
serve a niche best supported by closed methods. As the sale value
of software diminishes, the use value will rise for everyone. This
in turn will raise the value of the skills necessary to make best
use of the growing open software pool.
Regardless of the accuracy of my conclusions, one thing is
clear. Open Source software will continue to exist as long as its
user base wants it to. Given the growth in this user base today,
any assumption that Open Source software will fail on its own
merits is not well supported at all.
Respectfully,
Doug Dingus
Related Story:
Financial
Times: Why Open Source is Unsustainable(Oct 23, 2004)