Several mechanisms are available for interacting with the ISNI database, including manual interaction, batch processing and real-time API requests and responses.
Individual users of the system, including the general public, can use the standard search facilities via the Search web page of the ISNI website. More extensive search and other facilities are available to ISNI Members and Registration Agencies, using either a web interface or a “WinIBW” facility supported by OCLC.
Batch processing of data is also utilized, for initial or bulk loads from ISNI Registration Agencies and Members, for matching purposes and particularly for ISNI assignment requests. Because of the potential variability between sources in data quality and richness, there can be significant interaction between the data contributor concerned and the ISNI Assignment Agency before a live load takes place. There may be some iteration, involving requests to modify or tweak incoming data structures, and the ISNI Quality Team may decide to sample potential batch loads in a test environment so as to keep quality standards high.
Real-time API interaction with the ISNI database is increasingly used, in order to automate processes as far as possible, help synchronize databases and eliminate turn-around delays. Two main APIs are currently deployed: an “SRU Search” API and an “AtomPub Assignment Request” API. Both are extensively documented in the Technical Documentation section of the ISNI website.
The existing APIs already provide a number of features, filters and options. Some discussions are underway to establish whether the functions supported could usefully be extended, for example to use the AtomPub API for automated updates to existing records and other operations. ISNI’s Technical Advisory Committee will review these and other suggestions before making recommendations to the ISNI-IA Board.
ISNI supports a number of data formats, both for submission to the database and for outputs from the system.
For data input operations, primarily related to searches, to matching between source databases and ISNI, and to ISNI assignment requests, ISNI-XML is the preferred format. It is also possible to submit tab-delimited CSV information for batch assignment requests and matching.
For data output from the ISNI system, ISNI-XML is again the preferred and richest format available. Additionally, as a product of ISNI’s work on Linked Data , it is also possible to access a subset of the data available in either RDF/XML or JSON-LD formats.
For more details go to the Technical Documentation section of the ISNI website.
Beyond these standard offerings, some work has been done on establishing mappings between ISNI-XML and metadata formats widely used in particular sectors. Two examples of this are MARC21 and ONIX. The eventual aim is to evaluate whether it might be possible to accept inputs to ISNI in these or other formats that are “native” to particular constituencies of users, and thus to lower barriers to adoption and uptake.