Abstract
Persistent Identifiers (PIDs) have lately received a lot of attention from scientific infrastructure projects and communities that aim to employ them for management of massive amounts of research data and metadata objects. Such usage scenarios, however, require additional facilities to enable automated data management with PIDs. In this article, we present a conceptual framework that is based on the idea of using common abstract data types (ADTs) in combination with PIDs. This provides a well-defined interface layer that abstracts from both underlying PID systems and higher-level applications. Our practical implementation is based on the Handle System, yet the fundamental concept of PID-based ADTs is transferable to other infrastructures, and it is well suited to achieve interoperability between them.