This ISPS retrospective begins with the central problem facing modern research: too many published findings cannot be independently reproduced, undermining trust in science. In response, Yale’s ISPS Data Archive has spent the past decade building systems to ensure that data, code, and documentation are preserved, curated, and fully usable by others. The article highlights the ongoing challenges — from incomplete metadata to code that won’t run outside the original environment — and explains why reproducibility requires more than openness; it demands standards, infrastructure, and active stewardship.

The piece also notes encouraging progress: more journals require data sharing, more researchers embrace version-controlled workflows, and FAIR principles are becoming the norm rather than the exception. Still, ISPS argues that solving the reproducibility crisis depends on long-term collaboration among researchers, curators, and publishers to ensure that scientific claims can actually be verified.