The workflow in ISNI assignment and diffusion is a scrupulous process designed for maximum accuracy. The quality of ISNI metadata is paramount, with all ISNI participants playing a role. Data contributors and Registration Agencies are responsible for the quality and completeness of the input data and also for taking action on feedback reports provided by ISNI.
The ISNI Assignment Agency runs the matching algorithms at the core of the ISNI system and maintains a suite of programs that sample data, check for anomalies and enrich records. The role played by the ISNI Quality Team at the Bibliothèque nationale de France and the British Library is crucial. This team evaluates new data sources, identifies strategic databases to add, reviews statistical samples, and corrects records flagged for manual review, as well as responding to end user input.
The ISNI database has been built from contributed files from many sources across multiple domains. In this way the ISNI core data is enriched, assignments are corroborated and confidence levels are increased with each new source. And at the same time ISNIs are diffused widely as enriched metadata is fed back to the contributing source. Data may be loaded to the ISNI system via batch load or online using web page forms or an assignment request API. Unlike online requests where ISNI assignment is always achieved, an ISNI is only assigned for a record contributed via batch load where there is a confident match with an existing record or where the name is unique and does not exist in a shorter form. The ISNI Quality Team samples each new source and may recommend single source assignment for strategic files in new domains.
Once ISNIs are assigned, they are made freely available along with their associated metadata. The database is accessible from this web site and also via a search API. Data contributors receive reports after each batch load and they also receive periodic notifications including new assignments for records that previously had only provisional status. Also in the notifications are indications of additional new sources on records and error corrections. Data contributors are encouraged to use and rediffuse ISNIs widely. See also: Linked Data
There are two methods of online input, via web page forms and via the Atom Pub API. The web interface allows data contributors to make corrections and enrichments to existing records, to add source codes and local identifiers to existing records and also to add new records. There are two different templates, one for persons and one for organizations. The Atom Pub API enables ISNI Members and Registration Agencies to submit requests for ISNIs using their own interfaces. The API uses the ISNI request schema.
The public database includes a prominent invitation to all users to submit enrichments, clarifications and corrections to the information held by ISNI. Captcha is used to avoid machine input. Inclusion of an email address for a response is optional. The data is stored in two non-display fields, one a general note and the other a web page reference (URL). These fields are indexed and a daily robot generates an alert email that is sent to the ISNI Quality Team for follow-up. The notes and URLs may be converted into public fields during the review process.