In brief, Edifix is able to find the Crossref DOI for your reference if (a) Edifix can successfully parse your reference, (b) the cited reference is indexed by Crossref, and (c) the Crossref record for the cited material contains enough corresponding data.

Assuming that (a) and (b) are both true, (c) can mean different things for different types of references. The matching algorithms that Edifix uses to find and insert Crossref DOIs differ for book and journal references.


For books, Edifix uses the author names, book title, and year of publication to search for the corresponding record in Crossref. The publication location and publisher components are not used in our search algorithms because Crossref’s internal query system doesn’t use these fields. Because of the limited amount of data used to make these matches, the author, title, and year metadata must match exactly the data in the Crossref record in order for Edifix to (a) safely find the unique corresponding record and (b) insert the correct DOI.

These matching requirements are the same for any system making metadata and linking queries to Crossref, and so it is imperative for publishers to submit full and accurate book metadata to Crossref.

Even if a book reference is fully parsed and restructured by Edifix, it still may not link because of the available data in Crossref. To see what data Crossref has available for your reference, go to and search for the referenced material (or the DOI, if you know it already). Within the Crossref record for the material, select Actions > Cite to view the data Crossref has for that record.


For journal references, the “fuzzy matching” algorithms are more sophisticated and can handle “looser” input. In addition to the authors and publication title, Edifix uses the journal title, volume, issue, page, and date information to make the match with Crossref. Because journal references include more components that we can match on, Edifix is more lenient with mismatching article title and author information. Some of the metadata can be missing or incorrect in the reference if the majority of the other components are exact matches with the Crossref record.