newheader

Free Software in Science

There have been numerous discussions on the parallels between science and Free Software in the past. In fact indications on how science and Free Software operate on similar principles can be found regularly.

Fundamentally, science and Free Software have in common that both build upon cooperation of many people that in cooperation achieve more than each of them could have achieved alone. The best quote describing this is by Sir Isaac Newton, who said: If I have seen further it is by standing on the shoulders of Giants.

The advantages of cooperation do not only benefit the scientists or developers involved, but society as a whole — and the notion of society here does include economy. It is characterizing that the advantages are also available to those who have not contributed to them or maybe even fought against them actively (just consider the people who fought the notion that earth was not a disk).

Within these parameters, the connections are pretty clear to most people. But there is yet another, hardly understood connection, resulting from the scientific method.

Part of the scientific method is to create new theories and conduct experiments that will verify them. Has a theory been experimentally verified once, any additional verification will not add to the scientific knowledge base.

In other words: By having been experimentally verified, a theorem is considered valid. Additional verifications will not make it “more true than true.”

Another case is disproving a theory by conducting an experiment that contradicts the theory, falsifying it. In that case, the theory in its current formulation is not true. It needs to be modified or discarded alltogether — independent of the amount of verifications that existed for it.

One single falsification can make an unlimited amount of verifications irrelevant. Falsification is an essential part of the scientific process. Without falsification there is no science.

Where is the connection to Free Software?

Software becomes a more intrinsic part of science, which does not refer to using text processing tools to write down the results. From the viewpoint of scientific quality it is irrelevant whether a scientist is using plain ASCII text files or a certain word processor for their publication.

Although in the latter case it is probable that in a few years time the results will have to be entered anew because the current program cannot read the old file properly or maybe not even at all.

The connection between software and the scientific method exists when experiments partially or entirely rely on software. This means the software becomes part of the scientific process and result.

And to anyone who has ever developed software it is obvious that knowing an algorithm is not sufficient to provide means for falsification; implementation is equally important and also becomes part of the scientific result.

Proprietary software by nature creates a “black box”. If you wish, you can visualize it as a small, black box with a button and a light. Now someone tells you that when you press that button, a certain experiment is going to be conducted and when the lamp lights up, it was successful.

One could raise the question what knowledge a person could gain From pushing the button and seeing the lamp lighting up.

It gets really fascinating, when a second person comes with a second box, which claims that the same experiment is going to be conducted — but this time the lamp does not light up when pushing the button.

In none of these cases do we have the capabilities to verify or falsify, everything is based upon belief and trust only. This implies a very interesting conclusion.

Proprietary software is incompatible with the scientific method!

Additionally, we are experiencing other problems. Science is not only about what we know today, but also how we got there. The way we took, the way a discipline developed, is part of the cultural heritage of mankind and may contain important information for future generations. Results must not be time dependent.

Should anyone wish to repeat an experiment of Leonardo da Vinci today, this can be done. It may require work, but it is possible.

Making the assumption that the software used in an experiment still exists and did not — as usual — get destroyed because after 10 years the last remaining floppy disk containing the program got thrown in the trash. What are the chances for an experiment based upon proprietary software that people will be able to find the specific hard- and software versions that the software depends upon?

That probability is obviously reaching zero pretty quickly, especially when thinking in timescales of generations.

Free Software with its freedom to port it to other platforms really does allow to repeat such experiments and the good or not-so-good ideas people had.

Therefore, Free Software helps archiving the scientific and cultural development of mankind. It allows preserving the way of how we got where we are today.

So when going into detail, it becomes apparent that the connection between Free Software and science is much more intensive than it seemed at first. Also this shows the connection to social and cultural aspects that connect mankind.

Copyright (C) 2003 Georg C. F. Greve <greve@gnu.org>

Permission is granted to make and distribute verbatim copies of this transcript as long as the copyright and this permission notice appear.