Understanding Free Software: Research questionsAlan Cox, Linux/Red Hat, UK
Against a background of nay-sayers and the beginnings of a deep recession free software has flourished. Many predicted that the Linux companies mixing proprietary software and free software would do the best. Most of those mixed companies never became public, those that did have little value and operate in a market dominated by purist free software players. It was predicted that the open source world would never have a good wordprocessor or user friendly desktop environment. They now do, to the point that proprietary vendors are ditching their proprietary Unix desktop environments for GNOME 2.
Europe stands at a point where service based software models are essential. It is hard to find a company that has not moved its actual bulk programming work to the Czech republic, to the Phillipines or to India. Raw software creation is a relatively low value and labour intensive business. It has as little future as an effective profit-making business model within the EU as farming. Possibly it has less because software is much easier to transport and does not decay or suffer from quarantine rules. An open source, and service based industry exists close to its customers and makes much more efficient use of programming time by continually refining and reusing existing tools.
Understanding the open source movement has been hard, in part because of the lack of good research done by people who understand the community (as opposed to those trying to force fit it into their convenient existing models). A second problem arises because the most well known studies from inside the community lack academic rigour and in at least one case come from an extreme political viewpoint which denies the existence of society as a concept.
The models put forward by the scientists consistantly break down. They predicted a shortage of talent while the community was grappling with questions about how to grow at high rates. They misunderstood the nature of computing people and graphical desktop environments. Even today they misunderstand the technical computing worldview of wordprocessing tools. The models put forward by the free software community lack academic rigour and do not appear to work much better.
What exactly are the things that make free software work ? This is the first thing we have to understand better and we must do such research ensuring they are done with academic rigour and including people who truly understand the open source model.
The second question arises in part from the first. Having seen how effective the model is with software it becomes important to ask how relevant it is elsewhere. The computing world is unusual in that it deals almost entirely with intangible objects. Computer software is like a book or a painting, and indeed good programmers can see beauty and structure in code the same way people see them in a poem. Computing has a very low entry cost, creating a large pool of opportunity to contribute, and a very effective mass communication system making collaboration close to zero cost.
There are many obvious parallels between computing and literary works. One however in most cases seems to be a real barrier. With the exception of things like encyclopedia and dictionaries a literary work is not normally divisible between many authors to give good results. In the dictionary case SLUG are already compiling dictionaries of new words and using the same models. Freeddb also uses similar models to compile data on music CDs and their contents.
Understanding where else the open source model can be used could help in cutting costs and in improving many systems within the EU. The open source world has already demonstrated it can be an effective way of localising services, providing the data that is required is available in an easy to process form, and the rights to the data do not conflict with localisation.
It is important that we begin to understand where the model works, and where it breaks down in order to improve the efficiency of the EU, and the businesses within.