8 décembre 2014
Bernard Bel et al., « Deterministic assignment of PIDs : why and how ? The SLDR experience », HAL-SHS : sciences de l'information, de la communication et des bibliothèques, ID : 10670/1.lnul7g
In many digital repositories, persistent identifiers (PIDs) are produced with the help of a mechanism that makes them look like random strings. It is often argued that the PID should not display a trace of 'semantic' information related with the targetted document. This claim is challenged by expectations of producers of documents disseminated and preserved in the SLDR/ORTOLANG (http://sldr.org) repository: producers wish to anticipate the creation of persistent links to files when these shall be contained in the package submitted to the repository and/or used by external query/processing devices. In this case, PID calculation should be deterministic and based on comprehensive syntactic rules. A few cases will be discussed. In the second part of the session, an algorithm for deterministically assigning PIDs to all documents contained in an archival item will be informally introduced, along with a minimum set of options which proved beneficial for a producer/user- friendly implementation of this practice.