The development process used to develop DHIS 2 is more like many other larger scale FLOSS projects. With distributed development between more or less independent actors. For this to work a community has to be built around DHIS 2. In this section I am going to describe how this is done for DHIS 2.
The HISP community is the most important asset for the development of DHIS 2. There are many different peoples within this community (or network if you will) ranging from ICT professionals to health workers, and with people from both high-cost and low-cost countries. HISP has close ties to the academic community and a lot of contribution is sought from this community. Contribution is also sought through paying ICT professionals in low-income countries. Through seeking contributions from this two sources an application can be built with very little monetary cost. Master students to not get any pay. PhD students gets their pay from the state or other sources of scholarship. ICT professionals from low-income countries like Vietnam to not demand much money.
At the University of Oslo, which is the central node for DHIS 2 development, a course is held called Open Source Software development (INF5750). The practical assignments of this course is to participate in the DHIS 2 development. Most of the students taking the course are not seasoned Java programmers, and the student spend a lot of time familiarising themselves with the frameworks used in the DHIS 2 development. The course is credited for one third of a semester, so there is a limit to what the students are able to do. Some of the students might catch a more long term interest and continue after the completion of the course, through writing a master thesis relevant for DHIS 2 and later on a voluntary basis. This is not unlike other project where there is a relatively small core and many who make small contributions. In the following paragraphs I will look into who the major contributors to the DHIS 2 core are.
Judging by the number of commits to the DHIS 2 Subversion repository there are three people who have contributed more than any others. All this people are Norwegian and are or has recently been master students at the University of Oslo. The three top commiters according to the number of commits are responsible for 50% of the total number of commits. Looking at the number of commits is a poor metric for measuring contributions, one commit can possible be more significant than a hundred commits.
To be better able to judge the value of the contributions I will use the size of the difference caused by a commit as a metric. When changes are committed to the repository the difference between the current version and the committed version is stored. The size of the difference is the number of characters in the diff log. This is a quite good metric because the addition or change of text files create large diff logs, but addition or change of binary files creates small diff logs. Source files are typically text files. By this metric only one of the three previous top committers are still among the top three. They are still among the top six committers, however. Among the new three top committers there are two Norwegian master students and one Vietnamese. According to this metric there are six committers that stands out. The six top committers stands for 79% of the total size of all the diff logs. All of these except one are Norwegian master students. In total there are 73 committers. See Table 10.1.
There are of cause other ways to contribute to DHIS 2 than to commit code. Contribution of documentation and support is just as important, and so is the maintenance of the HISP web site. The efforts made to test and configure DHIS 2 out in the field is not shown by the Subversion commit statistics, neither. It seams clear, however, that it is mostly Norwegian master students who contribute to the DHIS 2 core. HISP have hired four developers in Vietnam, who previously was master students. One of these came in third place according to my Subversion commit statics. The managerial role in the project is taken by two PhD students. In addition to the work done on the DHIS 2 core there is work going on in India, Vietnam and Ethiopia to test and configure DHIS 2. DHIS 2 has very limited cooperation with other faculties at the University of Oslo and limited contributions from the health systems where DHIS is used.
For the development of DHIS 2 we have a web site1 and commonly used services like version control, wiki and issue management system. An overview of the collaborative services used can be seen in Table 10.2.
Communication is mostly done through the e-mail list. More direct communication through the use of IM, telephone and talking face to face is also quite common. It is difficult to track the direct communication, so it is difficult to say how much communication is done in this way. The Wiki and Weblog are only used by a few.
DHIS 2 is currently licensed under a permissive BSD style license. This is done in the hope that it will make DHIS 2 more acceptable in business settings, and widen the use of DHIS 2.