What are PIDs / Handles?
Digital objects are often affected by changes in their location or ownership, making it hard to reference them through web URLs as these are prone to break over longer time spans due to a wide number of causes. In scientific publishing, it is therefore not recommended to refer to any kind of material using URLs, as any future reader may have difficulties discovering relevant information in case the management of the respective web servers fails.
However, a lot of useful information is available in digital form, and there should be a safe way to reference such information. This is the gap that the concept of a Persistent Identifier (PID) aims to address: to provide a name that can be used to refer to a digital object and possibly retrieve it even through changes of object location or ownership. The concept of PIDs therefore relies on an act of explicit registration with a Naming Authority and making sure that the management processes of a repository such as ESGF cover such changes. As part of the registration, the identifier is associated with some minimal amount of information, including the object’s current location. If the object location changes, the information in the registry can be updated so that the binding between identifier and current location does not break.
In the case of scientific articles and datasets, the Digital Object Identifier (DOI) has been established as a PID, and all major publishers support its usage. Significant parts of data available through ESGF have received such DOIs through DataCite.
Technically, the DOI System is based on the Handle System, governed by the DONA Foundation. As such, DOIs are a specific brand of Handles that come with additional requirements on metadata to submit and curate. This specifically involves all metadata required for citation purposes, such as the names of authors and publishers, date of publication and so on. An important policy for articles or data that receive DOIs is that they are subject to long-term archival, meaning that both the identifier and the referenced object must remain available far into the future.
However, these policies are not enforced for all Handles, only for DOIs. In contrast to a DOI, a Handle does not guarantee that the referenced data object or article remains available over a defined timespan; yet, the Handle itself will always remain valid and lead to at least a page displaying the information submitted upon the identifier’s registration. This allows ESGF to offer Handles as PIDs for data carts and preliminary data products, where the conservation status is yet unknown, preventing the use of DOIs, but also not indicating any guarantee that such data or cart contents will be kept.