DOIs Explained: Crossref, DataCite, Deposit & Resolve

The Complete Guide to Digital Object Identifiers, Crossref & DataCite

Each day, millions of researchers follow links to journal articles, datasets and reports without a thought for the hidden infrastructure that keeps those links dependable years, even decades, after publication. That infrastructure is the Digital Object Identifier system, and anyone who publishes or cites academic research today needs to understand how it works, who runs it and how to use it properly. This guide covers the whole picture: what a DOI really is, why the system exists, how Crossref and DataCite differ, how deposit and resolution work technically, and what all of it means in practice for authors publishing with a journal like JMVRA.

A Short History of the DOI System

Publishers in the mid-1990s came up with the DOI system after noticing a growing problem: as scholarly content went online, links to articles kept breaking. Publishers redesigned their sites, switched content management systems, merged with other publishers or closed platforms, and each time the URLs cited in thousands of other papers stopped working. This problem is called "link rot", and by the late 1990s it seriously threatened the integrity of the scholarly citation record.

To solve this, the International DOI Foundation (IDF) was set up in 1998 with the aim of creating a permanent identifier layer above any individual website. Authors would cite a DOI, a string that never changes, rather than a URL, and the DOI system would take care of always pointing to the place where the content currently lived. Later the DOI system was formalized as an international standard, ISO 26324:2012, which made it the accepted worldwide mechanism for persistently identifying digital content.

Why Not Simply Use an Ordinary Web Link?

An ordinary URL says where something is, while a DOI says what something is and leaves the system to work out its current location. At scale, the difference is enormous. Imagine a journal that moves from one hosting platform to another after ten years, so that every article's URL may change. If ten thousand other papers had cited the old URLs directly, all of those citations would break at the same moment. With a DOI nothing breaks: the publisher updates the DOI's target URL once, in a single place, and every citation to that DOI, old or new, keeps resolving correctly.

Who Really Issues DOIs?

Authors cannot register a DOI of their own straight with the International DOI Foundation. The IDF instead accredits a small set of Registration Agencies (RAs), each of which serves a certain community or type of content. For academic research the two most relevant are Crossref and DataCite, although others serve different sectors (for example mEDRA for European multimedia content, or Airiti for content in parts of Asia). A publisher, journal, university or repository becomes a member of the suitable Registration Agency, is given a unique prefix, and then registers DOIs for its own content under that prefix.

Crossref in Depth

Crossref, whose formal name is the Publishers International Linking Association, was established in 2000 by a group of scholarly publishers seeking a common reference-linking system, so that a reader could click a citation in one paper and arrive straight at the cited work, whichever publisher hosted it. That remains its defining strength: member publishers deposit reference lists along with their own article metadata, which lets Crossref assemble a huge citation graph linking scholarly works across publishers. Many citation-counting and bibliometric tools used in academia today rely on this foundation.

Crossref membership works at the organization level: a journal, publisher, university press or professional society joins directly, pays a fee that scales with how much it publishes, and receives one or more DOI prefixes. After that, the publisher must deposit accurate metadata for all of its output, including journal articles, conference proceedings, books, book chapters and, more and more, preprints and grant records. Accurate deposits matter because every downstream service reads the same record.

DataCite in Depth

DataCite came into being in 2009 to cover something Crossref was not built for: assigning persistent identifiers to research outputs that are not conventional publications. Datasets, software repositories, physical samples, instruments and even teaching resources increasingly required a permanent, citable identity of their own, distinct from yet connected to the papers that drew on them. DataCite designed its metadata schema and membership model around that need, and today it is the leading Registration Agency for university data repositories, national data centres and research infrastructure projects around the world.

One central DataCite idea is the related identifier, a metadata field that directly connects a dataset's DOI with the DOI of the paper describing or using it, and the reverse. A reader can therefore begin with a paper, follow a link to the data behind it, and rely on that link being a verified, machine-readable relationship instead of a passing mention in the text.

How the Handle System Connects Everything

Both Crossref and DataCite rest on a common layer of global infrastructure, the Handle System, run by the DONA Foundation. Technically, any DOI, whichever Registration Agency issued it, is one kind of "handle" recorded in this system. When someone opens https://doi.org/10.5555/12345678, the request reaches a proxy server run by the DOI Foundation; it asks the Handle System which URL is currently tied to that identifier and then sends the browser there with an HTTP redirect. Because this resolution layer is shared, a DOI minted by Crossref and one minted by DataCite behave the same for the reader, and nobody needs to know which agency issued a DOI in order to resolve it.

Which Metadata Is Really Stored

People often assume a DOI is "just a link." In fact each DOI record holds a structured metadata package next to the target URL, and that is why DOIs serve discovery and citation tools so well, beyond simple redirection. For a journal article registered with Crossref, the package normally covers the article title, the complete author list (frequently with ORCID iDs), journal name and ISSN, volume and issue, page numbers, publication date, abstract, funding details and licence terms. A dataset registered with DataCite instead centres on creators, publisher, publication year, resource type and, crucially, related-identifier links back to associated papers. Since this metadata is standardized and machine-readable, indexing services, citation managers such as Zotero and Mendeley, and academic search engines can fetch accurate bibliographic details from a DOI alone, with no need to scrape or guess at what a web page contains.

Frequent Misunderstandings About DOIs

• "Having a DOI means my paper is peer-reviewed or of high quality." Incorrect. A DOI is a persistent identifier and says nothing about quality. Predatory publishers are able to register DOIs for their content, and some do.
• "As an individual author I can register a DOI myself." Not directly: DOIs are registered by member organizations such as publishers, journals and repositories, rather than by individual researchers, although your institution's repository might provide this for datasets.
• "A DOI is more or less the same as a URL." Not exactly. A URL names one specific location, whereas a DOI leads to whichever location is currently registered for it, and that can change over time while the identifier stays intact.
• "DataCite only handles scientific data." Incorrect. DataCite handles a wide variety of non-traditional outputs, including software, physical samples and audiovisual material, not merely quantitative datasets.

Research Integrity and DOIs

Apart from being convenient, DOIs increasingly support research integrity and reproducibility. Journals and funders more often expect authors to deposit the underlying datasets or code under a DOI when they publish, allowing other researchers to check or extend the work. When a paper's Crossref DOI record points to its dataset's DataCite DOI record, the result is an auditable trail from published claims to supporting evidence, something ordinary hyperlinks inside a PDF could never reliably promise in the long run. Documenting your data and code clearly helps anyone who follows that trail to reuse them.

What This Means for Authors Who Publish with JMVRA

For an author publishing a research paper with JMVRA, the practical point is simple: a Crossref DOI gives the article a lasting, citable identity that keeps resolving properly even if JMVRA's own website or platform changes later. It improves how the paper is tracked in citation databases, makes it easier to list correctly on a CV, ORCID profile or institutional repository, and meets the expectation, growing at universities and funding bodies, that a "properly published" article has a resolvable DOI. JMVRA keeps this optional and priced separately at ₹500, instead of quietly folding it into the base fee, so each author can decide whether a particular paper needs it.