Archives are created in order to provide information about the past, even if it is the recent past. Not all information is captured in the first place and there are limits to how much information can be retained due to resource constraints. Future analysis of the information will depend on the nature of the original selection process and the information that is subsequently available for analysis. Retrospective science and scholarship concerns the reconstruction of the past using extant information. The effectiveness of future retrospection can be boosted by optimizing the quantity and quality of the information retained: avoiding unnecessary redundancy (through parametric curation) and by adaptively appraising and retaining the information that is most potent for retrospection (through adaptive introspection). The diversity and scale of the digital universe necessitates efficiency across the archival life cycle. The paper aims to advance digital curation theory, albeit using a practical prospect.