This talk present the past few years experience in promoting open-source for critical operating space systems. Space industry is highly reluctant to changes and favors tight control, so going to open-source is a challenge.
Space systems are extremely costly and long projects. They are difficult to validate and errors may imply catastrophic effects (up to complete loss of billions dollars missions). Space industry is therefore reluctant to changes and retain obsolete technology for years. Projects are also tightly controlled and each agency uses its own set of custom developed software stacks for tens of years. As funding has steadily decreased in this field as in all other industrial fields, new approaches had to be used: it was not sustainable anymore to maintain old stuff and meet requirements for new missions with higher demands. This talk presents the journey of an open-source project based on Apache Commons Math and its struggle for life from the early days to the acceptance as the basis for space flight dynamics software by space agencies. Getting the "Apache way" is difficult for the legacy space actors.