May 25, 2004

A model for open source software development

Author: Wilbur Pan, M.D., Ph.D.

Critics of open source software have argued that, from an economic standpoint, "giving the product away" makes no sense. As a new business model, open source has been the victim of many misconceptions and mischaracterizations: that it's the equivalent of communism, that's it's like a virus, and that it's economically dangerous to society, for example. In such a hostile atmosphere, it's not easy to objectively assess a new paradigm that differs greatly from the status quo. However, though a comparable model may not exist in the IT sector, a parallel working model for open source already exists in at least one other field, biomedical research, and this model has been proven to be of great benefit to society.

I am a pediatric oncologist, and I don't pretend to have any proficiency at programming or hacking. But to paraphrase our favorite paralegal open source advocate, research is what pediatric oncologists do. Throughout my career I have performed research from both a basic science standpoint and from a more direct clinical perspective. One of the basic underpinnings of biomedical research is the free and open sharing of information. We simply could not perform biomedical research without doing so.

For those unfamiliar with this field, here's how research gets done. First, a researcher defines a problem or question, and then designs an experiment to determine the answer to the problem or question. After the experiment is done, he or she publishes the results as a paper. The paper includes all of the details of the experiment, including enough details as to how the experiment was done so that another person would be able to duplicate the experiment. This paper also contains references to previously published research in the field. Finally, the paper is published in a peer-reviewed journal, which means that the paper is first reviewed by experts in the field. Followup experiments later refer to this paper, just as the paper in question refers to previously published work. Other researchers are free to take any published research and attempt to build on it.

The analogy to open source software development should be clear. For example, consider Writer. First, the developers chose a problem to solve (make a word processor that can edit Microsoft Word documents). They then wrote a program and made the source code available ( Writer beta release). The program was reviewed ( Bugzilla). Improvements were made, and the new software was published ( Writer final release). References were made to previous contributions to the field (documentation, contributor lists, release notes). And programmers are free to make contributions to the existing code or fork the program in a different direction ( Writer branches vs. trunk development).

If we look a little deeper, we can draw other analogies. In biomedical research there are companies that pursue research in the private sector (pharmaceutical companies), and the results of their research (drugs in development) are not immediately available to the research community as a whole. This is analogous to software companies that develop closed, proprietary software.

It has been argued that the only way to make money off of software is to follow the closed proprietary system of software development. If this were true, then no company would be able to make money in biomedical research, which depends on full disclosure and published research. It can hardly be argued that there is no money to be made in biomedical research. Pharmaceutical companies do make money. But they do so in no small part due to the fact that they participate in research that is published in peer-reviewed journals.

Suppose that Omnimed, a fictional pharmaceutical company, comes out with Gliomaway, a fictional new drug they think will help cure pediatric brain tumors. Omnimed could try to bring this drug to market by conducting closed trials and by keeping the properties of the drug hidden from view, but Omnimed would never be able to bring a viable product to market this way. No one in the medical community would use it. Instead, what Omnimed and many real pharmaceutical companies do is cooperate with people and institutions that participate in publishing peer-reviewed research. Through this process, every aspect of Gliomaway would be made available for evaluation: how it is made, its chemical structure, its mechanism of action, preclinical data obtained from cell culture and animal experiments, toxicity profiles, clinical trial data -- the list goes on and on.

In my field, I have a particular interest in pediatric brain tumors. By reading and keeping up with the latest published literature, I would become aware of Gliomaway, and if I thought it was worthwhile for my patients, I would approach Omnimed with a proposal to use Gliomaway in a clinical trial, with the understanding that the results of this trial would be published for everyone in the field to see. In most cases, Omnimed would provide the funds to carry out this research. In this way, it would be determined whether or not Gliomaway would be useful in the treatment of pediatric brain tumors.

Funding for biomedical research can also come from sources outside of the private sector. If I were able to develop a new drug for the treatment of pediatric brain tumors on my own, I could go to the NIH or some other governmental source for funding to develop this drug. There are also not-for-profit organizations that provide funding for biomedical research.

I can't emphasize this point enough: All of this work is carried out in the public domain. Anyone can examine the medical literature for himself -- all it takes is a trip to the nearest medical school library, or for those who are inclined to let a computer do the work, surfing over to

Now compare the current situation in biomedical research to the IT field. There are many companies that are creating software packages that others cannot examine. The source code is locked away. If a software company wants to make a product that works with another software product, usually companies reveal only the bare minimum of details to allow interoperability.

To my mind, what is happening with IBM, Novell, and the other big players with resources for funding the development of open source software is that they are taking the position of pharmaceutical companies in the biomedical field today. Novell and IBM are working with the open source community to develop products rather than trying to carry out software development completely under a closed model. The exact details as to how they will be able to generate profits remains to be worked out, but it can be done. Biomedical researchers, myself included, can easily sustain a living by "giving their product away" in peer-reviewed journals.

To be sure, the analogies between the IT field and biomedical engineering are not perfect. Pharmaceutical companies are able to take advantage of patents, but on the other hand, pharmaceutical companies arguably invest more money and resources for product development than a software company. Pharmaceutical companies need to pay for lab space, equipment, chemicals, and so forth, whereas a software company's main physical investment is in computers. The issue of software patents is a touchy one, to say the least, and is outside the scope of this article.

The point remains, however, is that biomedical research is part of a thriving industry that relies on the free and open exchange of information. It is impossible for a pharmaceutical company to bring a drug to market without publishing details about its product. The IT industry would do well to understand how other industries conduct business, and more importantly, how information is used in other fields. IT companies should pay particular attention to the issue of funding in biomedical research. Any biomedical researcher can apply for research funding from the public sector.

China, Japan, Korea, and other Asian countries have already expressed interest in encouraging Linux development., KDE, and Gnome have created not perfect, but very viable software products with a minimum of funding. Imagine what would happen if a full-blown desktop environment and office suite were created due to an investment in funding from one or all of these countries. Imagine how good these products would be with an active funding process.

Just as pharmaceutical companies work with biomedical researchers who publish peer-reviewed research for all to see, software companies should engage the open source software movement rather than be at cross purposes with it.

Wilbur Pan, M.D., Ph.D., is a pediatric hematologist/oncologist practicing at the Cancer Institute of New Jersey and the Bristol-Myers Squibb Children's Hospital in New Brunswick, NJ. He is Assistant Professor of Pediatrics on faculty at the University of Medicine and Dentistry of New Jersey/Robert Wood Johnson Medical School. He has been a user of Linux for the past 1 1/2 years after having had his Windows laptop crash on him one too many times, and has not had to deal with email viruses since.


  • Open Source
Click Here!