GINorden 2006

The call for participation to GINorden 2006 (PDF)

Ari Jolma will give a presentation under the tentative title: "International networking in the open source GI field".

Submitted abstract of the talk:

The “open source” is an approach to solve certain problems, which arise when information and ideas are considered property. Open source is based on the idea that greater common good is achieved when people can freely read, redistribute, and modify documents. Open source ideas have mostly been applied to computer programs, but they are increasingly been applied to other documents, data, and any digital information as well. The spreading of the idea of free software and documents coincides with that of the Internet and the web. The Internet has several roles for open source: (i) it is the main distribution medium for the free software and documents, (ii) it is a communication medium for developers and users of open source, and (iii) it hosts development platforms for open source developers. All these roles unite in web-based communities that have been borne around certain open source software and issues. A major step forward was taken in early 2006 with the founding of Open Source Geospatial Foundation (OSGeo). OSGeo was founded exploiting the model of Apache Software Foundation, a successful foundation that provides support for the Apache community of open source software projects. OSGeo comprises a board of nine directors, 45 voting members, software projects, and several committees. OSGeo functions on the resources of people who participate on its activities in the Internet such as mailing lists, internet relay chat (IRC), collaborative websites, and in the real life, such as meetings and the conference “Free and open source software geoinformatics”.

update: The paper has been accepted for presentation. Extended abstract is due August 30. The website of the Nordic GIS conference in Helsinki 2006

Extended abstract
OSGeo: International networking in the open source GI field

Keywords: free software, free geodata, community-led projects, web collaboration

Author and speaker: Jolma, A.

The “free and open source” is an approach to solve problems, which arise when expressions of information and ideas are considered property. It feels natural for many people who develop software to publish and share their work as free for other people to use and build on. They also think that a greater common good is achieved if ideas and their expressions as digital content and tools are freely distributed, modified, improved, and redistributed. The free and open source world relies on service economy, while product economy is concerned about buying and selling things. In this the free and open source world is in line with the current developments in industrialized economies, where the service sector has grown faster than manufacturing for a long time. The free and open source approach has traditionally been used in software development, but more lately interest in distributing documents, data, and other digital information freely and openly has grown. The most prominent promoter of free software development has been the GNU Project. The approach has been applied to publication of scientific literature by, e.g., PLoS, the public library of science.

Internet and the Web have been instrumental in spreading the idea of free and open source and in developing the software. The tools of the Internet, email, web pages, news, IRC, software repositories, Web-based development environments, and many others are the tools of the online communities. Online communities have their roots in the bulletin board systems, which preceded the Internet. Online communities and other societal effects of the Internet have been studied extensively. It is clear for example that the challenges an Internet community faces are not simply technical but also sociological: the challenges of social interaction and social organization (Kollock, 1996). Some Internet communities and their free software projects have been very successful. Eric Raymond (2000) tried to explain their success by what he called a ‘bazaar style’ model of software engineering. Raymond listed several key factors, which contributed to the success of Linux: attractive problem, personal characteristics of the community leaders, almost simultaneous invention of the Web, and so on. He also claims that the observed successes challenge many earlier results and assumptions like Brooks’ observation that adding developers to a late project makes it even more late (Brooks, 1975) and the tangible benefits utility function assumption of classical economics. Since its inception 1991, Linux has grown and matured and is currently a core element of several ‘distributions’. A distribution is a complete operating system, which combines Linux kernel with GNU software and other tools and applications. Many commercial companies are involved with the development of Linux and Linux-based distributions.

The field of geographic information is involved with geospatial data management, utilization, analysis, and modelling. The field uses software and information systems, such as geographic information systems, image processing software, and data access libraries, etc. The amount of existing geospatial data increases at an increasing rate, the main source being satellite remote sensing. More complex geospatial data sets are developed manually and they describe for example road and other networks, properties, buildings, and demographics. This data often has broad interest and uses. A practical problem has been that geographic information has been stored in many technical formats and even modelled multiple ways. The approach to distribution of data varies, a big difference can be observed between Europe, where data producers usually charge for data, and USA, where federal organizations are usually required to share data freely and openly. Few large companies dominate the software market for geospatial data management.

Geographic information processing needs of individuals and organizations have resulted in many research, specification and software development projects. The GRASS GIS, originally developed and released into public domain by the U.S. Army Construction Engineering Research Laboratories, is today a free and open source software project developed by a large international community. Scientific papers are usually openly available and they may be accompanied by free and open source software or they may inspire people to implement such software. For example some GRASS functions are implementations of algorithms published in scientific literature. The Open Geospatial Consortium, a non-profit, international organization has developed and released open and free-of-charge specifications for geospatial information processing. These specifications have inspired a large number of individuals and organizations to develop interoperable free and open source software. The problem created by the existence of many formats for geospatial data inspired Frank Warmerdam to initiate the GDAL project, which is a free and open source translator library for raster geospatial data. The possibilities that the Web created, and the initial lack of interest by the commercial GIS vendors on the platform, created an environment for a free and open source software and MapServer emerged as a successful product with a large user and developer community.

The OSG ’05 conference in Minneapolis, USA, turned out to be a landmark event. The OSG conference series is an outgrowth of MapServer Users Meetings that had started to attract free and open source geospatial software developers in general. Probably inspired by the community, people from Autodesk approached MapServer developers and other key persons in the free and open source geospatial world during and after the ’05 conference. They wanted to ‘open source’ their Autodesk MapGuide Web mapping software, which was one of the first Internet GIS and originally developed by a small Canadian company. They also wanted a foundation to be formed to take care of the new open source software project. After some behind-the-scenes operations, the proposal was published. The result was a heated discussion, some rethinking, and, eventually, establishment of the Open Source Geospatial Foundation (OSGeo) in February 2006.

OSGeo has a very well defined organizational structure. It is a corporation organized under Delaware law and has a complex governance structure. One reason for the complexity is an advance preparation for possible juridical problems. The main bodies of OSGeo are its board with nine members, committees, and software projects. OSGeo has currently seven committees and eight software projects, MapServer, MapGuide Open Source, GRASS, and GDAL among them. The committees develop and promote the OSGeo software suite, organize the operations of the foundation: fundraising, its website, its presence in meetings and trade shows etc., and think about the relations of the foundation and its software with the research and education communities. The foundation software projects are going through an incubation process, which looks at the quality of the software projects in general and at specific issues like copyrights etc. Developing the community is a much broader and more difficult process because the scope and span are much larger. The community is on one hand moving forward very rapidly, new ideas come up all the time on the email lists, on the wiki, on member’s blogs, etc. On the other hand, because of the strict structure, newness of many things, and because the foundation policy on many things has not yet been decided or settled yet, some things do not move so rapidly.

During the writing of this paper, an informal question: “What is working about OSGeo and what is not?” was raised on the community’s email list. From the discussion that followed, from other observations, and from personal experience it is clear that forming the foundation was a good idea. OSGeo, and its “brand” interest people and the members are interested in promoting it. OSGeo and its website also give one place for people to start looking for solutions. Main problems seem to be related to communication, there is a great interest but at the same time people are not always sure how to follow what is going on, how to contribute to it, and how the decision process happens. There may be a risk when this is combined with how people come up with ideas: seeing a good idea but not seeing it getting forward is disappointing and may turn people away. These kinds of observations are probably common among online communities, and the solution is perhaps simply that the people with talent for leadership exercise it and people who want to contribute, find a way to contribute in a way, which adds to the whole. At the same time, as Kollock (1996) points out, there is no algorithm that is guaranteed to result in a well working online community.

References

Brooks, F. 1975. The Mythical Man-Month: Essays on Software Engineering. Addison-Wesley.

Kollock, P. 1996. Design Principles for Online Communities. Harvard Conference on the Internet and Society. Also published in PC Update 15(5): 58-60. June 1998. Also available at

Raymond, E. 2000. The Cathedral and the Bazaar.