Astronomical software thrives on collaboration. Many collaborations, however, are created around funded projects and remain active only for as long as that funding continues. When the funding ends, the collaboration often ends with it, while the software and its users remain.

The consequences can be that software becomes unmaintained or former partners continue development independently, creating diverging forks which causes duplicated effort and uncertainty for new users and contributors. At the same time, the cross-institutional networks built during these projects disappear, networks that provide opportunities for discussion and grounds for future collaboration.

Community collaborations such as Astropy and the IVOA demonstrate that this is not inevitable. They show that sustainable collaboration can be achieved through a variety of ways of working and funding.

This BoF will explore what makes such collaborations last and why others fail. The session will be centered around a panel discussion with a panel of experts and the audience. We will explore questions including: What enables a collaboration to survive? Which governance and maintenance models work in practice? How should responsibility be shared between individuals and institutes? And which warning signs indicate that a collaboration is unlikely to remain sustainable?

Our aim, with this discussion, is to identify commonalities between successful projects, or what makes some initiatives less successful at remaining a long-lasting collabroration. If the outcome of the discussion permits, we intend to summarise those points and submit them as an entry in the proceedings, as a reference.