It is easy to show that the total contribution of a program to society is reduced by assigning an owner to it. Each potential user of the program, faced with the need to pay to use it, may choose to pay, or may forego use of the program. When a user chooses to pay, this is a zero-sum transfer of wealth between two parties. But each time someone chooses to forego use of the program, this harms that person without benefiting anyone. The sum of negative numbers and zeros must be negative.
But this does not reduce the amount of work it takes to develop the program. As a result, the efficiency of the whole process, in delivered user satisfaction per hour of work, is reduced.
This reflects a crucial difference between copies of programs and cars, chairs, or sandwiches. There is no copying machine for material objects outside of science fiction. But programs are easy to copy; anyone can produce as many copies as are wanted, with very little effort. This isn’t true for material objects because matter is conserved: each new copy has to be built from raw materials in the same way that the first copy was built.
With material objects, a disincentive to use them makes sense, because fewer objects bought means less raw material and work needed to make them. It’s true that there is usually also a startup cost, a development cost, which is spread over the production run. But as long as the marginal cost of production is significant, adding a share of the development cost does not make a qualitative difference. And it does not require restrictions on the freedom of ordinary users.
However, imposing a price on something that would otherwise be free is a qualitative change. A centrally imposed fee for software distribution becomes a powerful disincentive.
What’s more, central production as now practiced is inefficient even as a means of delivering copies of software. This system involves enclosing physical disks or tapes in superfluous packaging, shipping large numbers of them around the world, and storing them for sale. This cost is presented as an expense of doing business; in truth, it is part of the waste caused by having owners.
Damaging Social Cohesion
Suppose that both you and your neighbor would find it useful to run a certain program. In ethical concern for your neighbor, you should feel that proper handling of the situation will enable both of you to use it. A proposal to permit only one of you to use the program, while restraining the other, is divisive; neither you nor your neighbor should find it acceptable.
Signing a typical software license agreement means betraying your neighbor: “I promise to deprive my neighbor of this program so that I can have a copy for myself.” People who make such choices feel internal psychological pressure to justify them, by downgrading the importance of helping one’s neighbors—thus public spirit suffers. This is psychosocial harm associated with the material harm of discouraging use of the program.
Many users unconsciously recognize the wrong of refusing to share, so they decide to ignore the licenses and laws, and share programs anyway. But they often feel guilty about doing so. They know that they must break the laws in order to be good neighbors, but they still consider the laws authoritative, and they conclude that being a good neighbor (which they are) is naughty or shameful. That is also a kind of psychosocial harm, but one can escape it by deciding that these licenses and laws have no moral force.
Programmers also suffer psychosocial harm knowing that many users will not be allowed to use their work. This leads to an attitude of cynicism or denial. A programmer may describe enthusiastically the work that he finds technically exciting; then when asked, “Will I be permitted to use it?” his face falls, and he admits the answer is no. To avoid feeling discouraged, he either ignores this fact most of the time or adopts a cynical stance designed to minimize the importance of it.
Since the age of Reagan, the greatest scarcity in the United States is not technical innovation, but rather the willingness to work together for the public good. It makes no sense to encourage the former at the expense of the latter.
Obstructing Custom Adaptation of Programs
The second level of material harm is the inability to adapt programs. The ease of modification of software is one of its great advantages over older technology. But most commercially available software isn’t available for modification, even after you buy it. It’s available for you to take it or leave it, as a black box—that is all.
A program that you can run consists of a series of numbers whose meaning is obscure. No one, not even a good programmer, can easily change the numbers to make the program do something different.
Programmers normally work with the “source code” for a program, which is written in a programming language such as Fortran or C. It uses names to designate the data being used and the parts of the program, and it represents operations with symbols such as ‘+’ for addition and ‘-’ for subtraction. It is designed to help programmers read and change programs. Here is an example; a program to calculate the distance between two points in a plane:
float
distance (p0, p1)
struct point p0, p1;
{
float xdist = p1.x - p0.x;
float ydist = p1.y - p0.y;
return sqrt (xdist * xdist + ydist * ydist);
}
Precisely what that source code means is not the point; the point is that it looks like algebra, and a person who knows this programming language will find it meaningful and clear. By contrast, here is same program in executable form, on the computer I normally used when I wrote this:
1314258944 -232267772 -231844864 1634862
1411907592 -231844736 2159150 1420296208
-234880989 -234879837 -234879966 -232295424
1644167167 -3214848 1090581031 1962942495
572518958 -803143692 1314803317
Source code is useful (at least potentially) to every user of a program. But most users are not allowed to have copies of the source code. Usually the source code for a proprietary program is kept secret by the owner, lest anybody else learn something from it. Users receive only the files of incomprehensible numbers that the computer will execute. This means that only the program’s owner can change the program.
A friend once told me of working as a programmer in a bank for about six months, writing a program similar to something that was commercially available. She believed that if she could have gotten source code for that commercially available program, it could easily have been adapted to their needs. The bank was willing to pay for this, but was not permitted to—the source code was a secret. So she had to do six months of make-work, work that counts in the GNP but was actually waste.
The MIT Artificial Intelligence Lab (AI Lab) received a graphics printer as a gift from Xerox around 1977. It was run by free software to which we added many convenient features. For example, the software would notify a user immediately on completion of a print job. Whenever the printer had trouble, such as a paper jam or running out of paper, the software would immediately notify all users who had print jobs queued. These features facilitated smooth operation.
Later Xerox gave the AI Lab a newer, faster printer, one of the first laser printers. It was driven by proprietary software that ran in a separate dedicated computer, so we couldn’t add any of our favorite features. We could arrange to send a notification when a print job was sent to the dedicated computer, but not when the job was actually printed (and the delay was usually considerable). There was no way to find out when the job was actually printed; you could only guess. And no one was informed when there was a paper jam, so the printer often went for an hour without being fixed.