Four Free Software Fallacies
20 September 2002
Brendan is a lawyer specialising in IT and telecommunications
law. He welcomes comments on and criticisms of his papers.
Brendan's email address is email@example.com.
There are four fallacies relating to free software which seem to be getting
some press at the moment: (a) Giving stuff away doesn't work; (b) No
one will create free software applications; (c) The GPL is viral. and
(d) We need greed to drive developers. This paper addresses each in
turn and sets out why they are misdirected or otherwise do not form the basis
for an argument against free software when compared to the alternative proprietary
1. Giving stuff away doesn't work
This argument simply misunderstands the free software model. That model
attempts to create a market for software in which barriers to entry are minimal
and anyone who wants to compete can do so. However, it says that competition
should be services based, not product based. It achieves this by requiring
anyone who wants to compete to forego creating their own proprietary anti
competitive barriers to others entering the market. Nowhere does it
say that anyone has to work for free and no one is under any illusion that
you get something for nothing on the intial development. To repeat the cliche,
the free in free software is the "free as in speech, not free as in beer".
I would add here "free as in market, not free as in beer". What they
do say is that you do get something for nothing on the distribution side.
Welcome to the internet. They also say that users will the software
they need, because they need it. They'll be able to do it cost effectively
because they only have to innovate on existing software infrastructure, they
don't have to create that infrastructure from scratch. So, for example,
if IBM wants to pour money into commoditizing a complement for its products,
so be it. So be it also if some software isn't developed because no one wants
it. That's the beauty of the free market. Like it or hate it, it means
resources are used more efficiently.
2. No one will create free software applications
This is one of those "bumble bees can't fly" sort of arguments. The empirical
evidence is quite clear. There is a multitude of free software applications
available. There are more now than there were 12 months ago. Indeed,
free software has consistently gained ground over the last 10 years against
its rival proprietary model. It has gained that ground from more or less
a standing start against a model which had a 10-20 year lead and had (and
still has) very strong monopoly incentives. Free software applications will
be created where the market wants them. If the market doesn't want
them, there's nothing the proprietary model is going to do to save them (other
than prolong the agony by lock in and switching costs). Free software
critics need to stop deluding themselves. If free software was going
to die because it is a bad model (as opposed, for example, to coming under
political pressure from vested interests), it would be dead already.
3. The GPL is viral.
This is the most irksome of the four and is a fallacy which is oft repeated.
The GPL is no more "viral" than proprietary software. To the extent
it is viral it gives rights, it does not take them away. If someone
receives some free software they are not only permitted by the GPL to develop
it further, they are also permitted to use those modifications and to distribute
that development on the same basis that they received the source. What
is the position if they receive some proprietary source code? Prima
facie they are prohibited from doing anything at all to that code. If
they do make modifications to the code not only are they prohibited by law
from using or distributing those modifications, those modifications are vested
by the copyright law in the owner of the copyright in the source code.
If the GPL is viral, it is a set of viral permissions. It is
also a set of permissions of which the recipient is made aware of at the time
of acquisition (it's an obligation under the GPL to include a copy of the
licence with the distribution). On a similar analysis the copyright
law is a set of viral prohibitions and (potentially) takings. They're
a set of prohibitions which operate by default even in the absence of knowledge
by the recipient (by automatically taking copyright in the development work
from the developer and vesting it in the holder of copyright in the source
code). So in the proprietary case, not only can work created
by the developer be used by someone else, it can actually be taken from the
developer, and, on top of that, the copyright holder can sue them for breach.
As proprietary software companies are not required to inform recipients
of this, receiving "shared source" can be many times more dangerous to an
organisation than receiving free software. The "GPL is bad because
it's viral" is a tired argument which is well past its use by date. The same
arguments show that proprietary software is worse.
4. Software creation requires developers to be driven
The application of scarcity economics to public goods is at best fraught with
danger. The success of free software to date in the face of competition
from proprietary software must at least give us pause. The key assumption
underlying copyright policy is that development must necessarily be "big bang"
and be done in isolation by "entrepeneurs". It also assumes that it
involves significant sunk costs or capital outlay. Finally,
it assumes that sales, marketing and distribution is a serious problem which
needs to be overcome. On these assumptions market activity in relation
to software development will prima facie be seriously retarded. If
it is developed, it won't be distributed. As such, the Government gives
entrepeneurs monopoly rents in return for overcoming these barriers.
The fundamental assumption of free software is that innovation and development
comes from minor incremental improvement and massively scaled collaboration
and that distribution comes for free (for those of you who've been asleep
for the last decade, welcome to the information superhighway). If this
assumption is correct (and the fact that any, let alone significant, free
software resources exist with wide distribution is evidence that it is) it
means that productization is not necessary. If the free software movement
reaches its true potential, it will allow users to strip the monopoly rents
out of software pricing and put those rents to productive work elsewhere
in the economy.
The argument can sensibly be made that protection is an appropriate incentive
in markets where there is no opportunity to scale, such as specialised and
niche markets. However, to make it in relation to markets for commodity
items such as operating systems and productivity applications applies 17th
century thinking to a 21st century problem. If free software advocates
are proven correct, it will also be proof that in many cases copyright simply
sucks dollars out of productive areas of the economy and wastes them on monopoly
rents while simultaneously constraining innovation.
This is not a question about whether or not greed is an incentive. Rather,
it is a question about whether the greed game ought to be played on services
or whether productisation through proprietary software, with its anticompetitive
baggage, is necessary or desirable.
Of course, these criticisms rarely occur in isolation. They are regularly
attended by some call that business or government, or both, should shun free
software alternatives. Taken at their highest even these fallacies would
only sound a note of caution. Even in their absence proceeding with
caution is a wise course to follow. Neither business nor government
should follow the free software path if it's not appropriate to them. However,
that decision must be based on a sound and reasoned assessment of the
emprical data and one in which these fallacies should play no part.
The incentives of free software are exactly those at work in any other free
market. But there's the rub. The software industry isn't a free
market. It's certainly not as free as the free software movement could
make it. Barriers to entry and anti competitive features are so entrenched
as to be considered normal. We know that free markets produce more
for less, so why is there so much opposition to a free software market? Should
barriers to competition in the software industry be lowered? Of course
they should. Should Business take advantage of the opportunities that
the free software market creates? Of course it should. Does Government
have a part to play? Of course it does.