In this post I will introduce a talk given by Joseph Roumier (R&D Engineer at CETIC, the Centre of Excellence in Information and Communication Technologies, in Belgium), explaining the method that they followed in order to choose a suitable software forge to be installed for the CELLaVI project, which promotes adoption of Free Software by industry and institutions in Wallonia and Belgium.
They followed a three-phase process:
- First they pick up a list of the main known forge software. The list was too long (13 forges) so they needed to rule out most of them, and keep 4 (FusionForge 4.8, LibreSource 2.5, EmForge 0.28 and Redmine 0.8) in order to evaluate or compare them in a more exhaustive way. For this selection they applied strong exclusion criteria as not being libre software, or not in active development.
- To evaluate the 4 forges that passed this first test, they wrote the organization requirements (functional and non functional) and gave weights according of their importance or each requirement for the organization. Then they researched information to measure each requirement, from the public sources and also from their own user experience (installing the 4 forges in test machines). They also transformed the results of their evaluations in numerical data, in order to be comparable, giving scores (0 = absent, 1 = insufficient, 2 = good , 3 = excellent). Then they summed the weighted scores, and obtained a total score for each forge, resulting Redmine the best scored forge.
- They evaluated again the 4 forges, using the QSOS method for evaluating free software, and the result was again Redmine.
As the result of both evaluation processes was the same, they decided to install Redmine for their forge.