Automating Ingest of Metadata on Serials Subscriptions
This project was concerned with utilising the ONIX for Serials (Serials Online Holdings Version 1.0) format in an exchange of real data between a Publication Access Management Service (Serials Solutions) and a national union serials catalogue (SUNCAT) hosted by EDINA at the University of Edinburgh. ONIX for Serials is a family of XML formats for communicating information about serials products and Serials Online Holdings is the format for communicating serials holdings details.
The approach was to take files of information about serials holdings which had been created by Serials Solutions for two customers (Universities of Glasgow and Leeds) who were also SUNCAT Contributing Libraries, put the relevant information into SOH messages and transmit to EDINA. Representatives of the two organisations met in Edinburgh in November 2005 to discuss the content of the proposed files and the timing of the messages. The work involved in the creation of the files built upon the technical expertise developed by Serials Solutions during participation in the development of the standard. What was created was a flat A to Z list where each resource is listed separately with the hosted collection identified individually against each entry. The data was generated from normalised journal information (taken from fields in a MARC record in a defined order) or from vendor information when there was no available MARC record.
The two files were sent in November 2005. A further file was sent for Glasgow in January 2006 to improve synchronisation with the file sent to SUNCAT. Matching was carried out on the records from Glasgow and Leeds but the time taken to carry out matching was longer than anticipated and only Glasgow records were updated in SUNCAT. EDINA developed its technical capacity with SOH from scratch. The work involved parsing the message using a readily available tool (XML::Twig), extracting serials holding information and online service data and matching to SUNCAT records. The 856 field in the SUNCAT MARC records were then updated with the SOH data fields. Matching of incoming records with existing was most successful when records had both a title and an ISSN.
The key achievements were the development of capacities to create large SOH files by a PAMS and the development of a capacity to process and extract relevant data to update existing MARC records in a national union catalogue of serials. It is the view of both project partners that the development of technical capacities to handle the messages is not a major undertaking.
The project also demonstrated that the SOH format was fit for purpose and possessed the required data elements to allow for a meaningful exchange. It was noted that all the necessary documentation (eg comprehensive key for available IDType values) was not available during the project. It was also suggested that problems over large file sizes could be reduced by consolidating some data fields and separating header from the holdings data. Finally it was proposed that values for more identifiers could be explicitly listed.
It is recommended that in seeking Library Management Systems and Electronic Resource Management Systems the requirement to be able to process ONIX for Serials messages should be requested in Requests for Proposals and also that publishers, aggregators and Publication Access Management Services (PAMS) should be encouraged to develop capacities to create messages.