Re: License Dilemma

[Alexander Terekhov]

John Hasler wrote:
[...]
> In the real world few inventions can be exploited while keeping them
> secret.

Uncle Hasler, uncle Hasler. See Shakespeare, Tudor and Jacobean, plus 
Saxony below.

http://www.charvolant.org/~doug/gpl/gpl.pdf

---------
3.1 Intellectual Property

Whatsoever then he removes out of the state that nature hath provided,
and left it in, he hath mixed his labor with, and joined to it something
that is his own, and thereby makes it his property.[15]

The implicit point of view contained in this essay is a Lockean one.
Producing a piece of software requires taking the state of nature, the
common heritage of software tools and techniques, and using them to
fashion something new.

To the extent that programming involves labor — and thinking is
certainly labor, ask any student — a piece of software is [intellectual]
property. To the extent that invention requires labor, an invention is
property. This state of affairs is recognized in intellectual property
law, such as copyright and patent law.

Nothing, of course, prevents the creator from choosing to place the
fruits of their labor in the public domain, for whatever reason they
choose. Or to place it under the GPL. But the choice to do so is theirs.

3.2 The Economic Arguments of the FSF

It is worth examining the economic arguments put forward by the FSF.
These arguments are largely intended to refute any benefit that non-free
software production might have on society at large. Unsurprisingly, I
find these arguments rather unconvincing.

In one article, economic arguments are crudely mischaracterized as a
form of holding-to-ransom.[28] This article claims that the economic
argument is constructed in terms of proprietary software versus no
software at all. However, the economic argument is largely concerned
with notions of allocation of resources. It’s more efficient, in terms
of producing software that users want to use, to have a feedback loop
connected to users (more on this in section 3.3.2). Specialization and
division of labor suggests that allowing professionalization will lead
to a greater output of higher quality software; the FSF allows
professionalization, but only in a restricted environment. Ricardo’s law
suggests that allowing professionals to specialize and trade will
provide a greater total output of all economic goods, not just
software.[24]

Another article cites the ready copyability of software as providing a
different economic model to book publishing.[29] This argument entirely
ignores the underlying costs of producing a book and considers only the
costs of producing a hard copy; it also ignores the existence of the
photocopier. The major investment in any creative work, be it a program,
book or piece of music, lies the in the process of creation. A book
needs to be written, edited, re-written, typeset, published. A program
needs to be written, debugged, packaged. All of these things involve
work; this is where the notion of ownership comes from, not the
duplication of the final piece.

Finally, there is the question of wasted effort. An example is given of
closed software requiring wasted make-work to provide a suitably adapted
version.[28] This is a reasonable criticism of closed software, although
not of all proprietary software (see section 3.4). However, if make-work
is regarded as a waste, then the GPL forces make-work for anybody
unwilling to accept the terms of the GPL; this aim is implicit in the
argument for GPLing libraries.[30] If closed software imposes a social
and economic cost, then so does the GPL.

3.3 Amateurs, Professionals and Patronage

The main benefit to society and the economy of such intellectual
property notions as copyright and patents is the creation or
encouragement of a class of professionals. Allowing somebody talented in
a certain field to make a living directly from that field has a number
of advantages: the most obvious is the ability of talent to concentrate
on what they are good at, rather than requiring them to undertake other
tasks to support themselves; additionally, specialization is permitted,
leading to a feedback loop where skills are honed and improved. Prior to
copyright, those wishing to be inventors, authors or other creative
artists had to either find a patron or have additional means.

3.3.1 Patronage

In the past, anybody not of independent means who wanted to create
something intended for general release needed to find a patron: a person
willing to foot the bills in exchange for some intangible return. Many
creator-patron relationships were very fruitful, with the patron acting
as a source of inspiration to the creator. However, unless they were
beings of considerable foresight, patrons intended to use their
association with artists and scientists to further their own political
ends.

Patronage is, today, a little more sophisticated. The general model is
endowments to a university or other institution which, in turn, makes
research and thought available to the public domain. The patrons derive
advertising and public relations benefits from their sponsorship. Most
governments recognize the benefits to be had from a stream of new ideas
and research moving into the public domain and provide some patronage
through funding arrangements.

This form of institute-based patronage seems to be the ideal for the
FSF; a community of public-spirited developers all working towards the
common good.

The first problem with this model is the restricted financing available.
Endowments from commerce essentially come out of a public relations
budget; necessarily limited. Support from the public purse runs into the
political problem of taxation levels and control. In theory, all
software could be publicly funded by using the sums spent on commercial
software. The rise in taxation, even if it were to take less than the
cost of commercial software would be politically unpopular. The lack of
control over what software is delivered is also likely to be a point of
resistance.

The second problem with this model is that of feedback. Microsoft at
least tries to find out what their customers want—even if it results in
obvious idiocies, such as talking paper-clips.[26] The institute-based
model has no feedback mechanism— or, even worse, a feedback mechanism
based on internal squabbles or the political aims of the patrons. The
result is that the software produced tends to reflect the interests of
the programmers, rather than the users. This is obvious in most
open-source: system programs, utilities and development tools abound;
applications—with the exception of every hacker’s favorite, games—are
harder to find. The GNU offerings are almost exclusively system and
development tools.

The open-source movement has provided the impetus for another form of
patronage. Companies such as RedHat or Linuxcare need free software to
succeed to be successful themselves. As a result, these companies hire
the producers of free software to ensure the supply, provide good public
relations and provide in-house expertise for the support operations that
make up the companies income. This is in addition to the general desire
on the company’s part to do the right thing; open-source is still a
social movement.

Eric Raymond argues that this form of patronage works, in part, because
the companies dispensing the patronage are leaders in the field, and
thus benefit in proportion.[20] If this is true, then it also represents
the break-point for this form of patronage. As the market becomes more
competitive, a significant free-rider problem appears: companies that do
not have the overhead of patronage and can offer the same services at
reduced cost.[13]

3.3.2 Amateurs

The alternative to professionalism or patronage is amateurism: doing
something for the sake of interest in doing it.

The word “amateur” has acquired a negative connotation over the years,
having overtones of “halfbaked” or “poorly done”. This extra baggage is
unfortunate; many amateurs are of the highest levels of skill and
dedication,5 and I don’t intend to suggest otherwise. A notable feature
of amateurs is that they can approach problems with an eclectic
viewpoint which may be absent from a purely professional approach.

However, an amateur production of anything, whether it be a software
package or theater production requires either independent means or
another source of income and a dedication to using one’s spare time in
the pursuit of the production to the exclusion of other activities.6

Open-source software is the beneficiary of a peculiar state in the
software industry: many of those who are amateur programmers by night
are professional programmers by day. Linux—and the Unix approach, in
general — has made amateur programming attractive. Rather than produce a
large, complex program requiring a huge team to produce and maintain,
small packages can be produced. Software can easily be divided into
front- and back-end parts, so that functionality can be produced and
tested, without the overhead of a GUI.

The upshot is that open-source, at present, gains the benefits of both
amateur enthusiasm and inventiveness and professional knowledge and
discipline (and income). This blessed state of affairs exists while
there is a pool of professional programmers able and willing to use
their spare time to produce open-source software. I would suggest that
the aims of the FSF will reduce this pool enormously, and the effects
will be catastrophic. Eric Raymond has argued that open-source culture
is essentially a gift culture; resources are in abundance and you gain
status by the bestowing of gifts on the community.[22] The absence of a
large supply of well-paying professional jobs in software—more or less
predicated on a large scale commercial industry— will re-introduce the
economics of scarcity to the software culture.

3.4 Patents Are Your Friends

Open-source software has an enviable reputation for reliability. The
usual reasons given for this reliability are that the presence of the
source code allows immediate analysis and rectification of any problem
and that the ability to contribute enhancements and bug-fixes vastly
expands the number of people working on and contributing to a piece of
code.[21]

Generally, the notions of open-source and non-proprietary software are
conflated — as opposed to closed, proprietary software. However, there
is no particular reason to do so. Instead, software can be categorized
using two axes: an open-closed axis and a free-proprietary one.[2] The
benefits accruing to open-source software are largely connected to the
open-closed axis. It is the making source code available that allows the
peer-review and correction feedback loop to take off.

Making software proprietary does stem the flow of contributions, of
course; nobody particularly wants to contribute to someone else’s profit
at their own cost. To offset this effect, open, proprietary software can
easily provide a renumeration model, offering payment or royalties for
contributions. 

[...]

To see the benefits of patents (and copyright) in maintaining
information, a historical example might help. In the era of Shakespeare,
a would-be publisher only had to get hold of a manuscript, by fair means
or foul, to be able to copy and print it. (The first publication of
Shakespeare’s sonnets followed this pattern.) Similarly, any acting
troupe that could gain access to a play could perform it. As a
consequence, play manuscripts were deliberately obscured and divided up,
with an actor being given just his lines, along with suitable cues. The
natural result of this is that many Tudor and Jacobean plays exist only
in fragmentary form; we only have the works of Shakespeare today because
he was well-regarded enough for a syndicate to track down most of his
work and publish it in the first folio.[23, 3]

As another historical example, the technique of porcelain manufacturing
was almost lost to Europe. Saxony’s attempts to maintain a monopoly on
porcelain production led to obsessive secrecy and the intriguing that
secrecy brings. Information was kept in the player’s heads and withheld
for political advantage.[12]

True, patents carry an additional piece of economic baggage over and
above open-source: they encourage8 inventors to work on, and profit
from, their inventions. From an economic perspective, this is a quid pro
quo for having come forward in the first place, rather than keeping the
invention a secret.

Secrecy may not seem very relevant to software. In the general scheme of
things, true secrecy is difficult to achieve, as something, a program or
a file format has be be available and is vulnerable to reverse
engineering. This represents an annoyance, but not a block to the
dissemination of information. To prevent dissemination, if the
discoverer of a new algorithm so wishes, he or she can, to coin a word,
bureauise the algorithm: provide it only through the services of a
company to which you submit information for processing. Bureauisation
can occur whenever the process is complex, but the results simple.
Combinatorial problems take this form, for example optimal path
computation, or the prime factoring of large numbers.9 Bureauisation
represents an obvious social and economic burden, yet it is a natural
consequence of no intellectual property protection.

Ultimately, patents provide the kind of legal protection needed to allow
VAS vendors to open their source. Without some sort of protection,
secrecy and obscurantism, with their costs, rule. It is possible to
argue that the costs of intellectual property outweigh the costs of no
such protection. But I think the verdict of history is against that
argument.
---------

regards,
alexander.