Author identity is highly important in INSPIRE. We manage this by clustering papers into author profiles based on ORCIDs or INSPIRE IDs, if they exist in the paper’s record or, failing that, on simple name string matching. While the ID-based approach works very well, string matching is far less precise and often leads to papers going into the wrong profiles.

INSPIRE records for eprints are created by a nightly download of the metadata (author names, title, abstract, etc.). These contain no author IDs, leading to the above mentioned problems. To overcome this, large collaborations such as ATLAS and CMS include authors.xml files with their submissions to arXiv. These files contain information on all authors, including their IDs and affiliations, allowing INSPIRE to place the papers into the correct author profiles.

Information and instructions on INSPIRE Collaboration Author List (i.e. the author.xml file) can be found here.