In almost all areas of science, the creation and maintenance of software play a pivotal role in bridging theoretical research with practical applications. However, the lifecycle of academic software often faces significant hurdles, from development to long-term sustainability. This talk delves into the inherent challenges of producing, maintaining, and extending academic software projects, drawing on the experience accumulated in 25 years of development of the deal.II library (www.dealii.org). I will focus on lessons learned from our development experience, and how these apply to most academic research projects, emphasizing the risk of software obsolescence and the critical need for dedicated maintenance efforts. We will discuss strategies to mitigate these risks, including fostering a culture that values software development as a scholarly endeavor, integrating software engineering principles into the academic curriculum, and adopting collaborative models that ensure continuity and innovation.