ISBN, ISSN, DOI: What Each One Actually Identifies

§ 01A Number Is a Question in Disguise

Every identifier answers a question, and the question matters as much as the number itself. The ISBN, the ISSN and the DOI each answer a different question — which is why the same piece of writing can, entirely correctly, carry all three at once, and why using one where you need another always ends in frustration.

None of these systems identifies "a work" in any simple sense. Each one identifies a specific thing about a work: a particular edition, a continuing publication, a citable object at a location. Grasping that distinction is the entire key to understanding when and why each system applies — and why a researcher asking a librarian for "the ISBN of a journal article" is, without quite knowing it, asking a question that has no answer.

§ 02ISBN: The Edition in Your Hand

The International Standard Book Number identifies a particular manifestation of a book-like work. Not the story, not the text, not the intellectual content — the specific edition and format. A novel published simultaneously as a hardback, a trade paperback and an ebook carries three different ISBNs, even though the words inside may be identical. Revise the text substantially enough to call it a new edition, and the publisher issues a new ISBN. Add a new foreword, change the publisher, correct enough errors to publish as a revised edition — new ISBN. The number is tied to the commercial product as it enters the supply chain, not to the creative work as the author conceived it.

This is entirely by design. ISBN was developed in the late 1960s from a British bookseller's numbering scheme and adopted as an international standard by the early 1970s. Its purpose was logistical: booksellers and libraries needed a machine-readable code that could identify exactly which item was being ordered, stocked or catalogued. A 10-digit number served until 2007, when the system migrated to 13 digits aligned with the EAN barcode family — a move that brought books into the same identification infrastructure as other retail goods.

The anatomy of an ISBN reflects its supply-chain origins: a prefix element, a registration group (country or language area), a registrant (the publisher), a title element, and a check digit. Publishers obtain ISBN ranges from their national ISBN agency; the identifier is assigned before the book is even printed, because it has to appear in advance cataloguing systems and on the barcode. This is a publishing industry infrastructure tool that happens to be publicly visible on the copyright page. It says nothing about who owns rights, nothing about where the content can be found online, and nothing that persists if the publisher lets the book go out of print. It simply answers: which exact product is this?

A consequence of this design is that not everything called a "book" has an ISBN — and some things that do may surprise you. Ebooks from major publishers carry ISBNs; ebooks uploaded directly through some self-publishing platforms may or may not, depending on whether the author obtains one independently. Open-access monographs increasingly carry ISBNs too, precisely because libraries use ISBN-based cataloguing systems regardless of price. And no ISBN is ever reused: even if a title goes permanently out of print, the number is retired, not recycled.

§ 03ISSN: The Ongoing Publication

Where ISBN says "this specific product," the International Standard Serial Number says "this ongoing publication." A journal, a magazine, a newspaper, an annual report series, a podcast, an email newsletter — anything published in successive parts under a continuing title qualifies as a serial, and serials get ISSNs.

The ISSN does not identify an individual issue, let alone an individual article. It identifies the title as an ongoing entity. A journal that has been running since 1953 carries a single ISSN that covers every issue ever published under that title. Change the title — even a minor change, in some cases — and the ISSN network typically issues a new number for the new title, linking the two in its database for continuity. This is why libraries can use ISSNs to manage journal subscriptions: the number is stable across time for a given title, regardless of which issue you actually want.

The ISSN does not identify an individual issue, let alone an individual article.

Print and online versions of the same journal conventionally carry separate ISSNs: one for the print, one for the online. This multiplicity is not a quirk but a feature — the two formats are distinct serials in supply-chain terms, even if the content is the same. The ISSN network, coordinated through CIEPS (the ISSN International Centre) in Paris, maintains the global registry. National centres handle registrations locally; the result is a two-level system that functions something like an international library for publication identity.

Individual articles sit below the level the ISSN was designed to reach. The journal article is where the ISSN hands off to a different kind of identifier entirely.

§ 04DOI: The Citable Object at a Persistent Location

The Digital Object Identifier answers a different question again: not what product is this, not what publication is this, but how do I cite and reliably find this specific object — now and in the future? A DOI is an actionable, persistent identifier: it resolves to a location (or a metadata record) and is designed to keep working even if the underlying URL changes. When a publisher migrates content to a new hosting platform, they update the DOI record rather than breaking every citation that has ever pointed to the article.

The DOI system is governed by the International DOI Foundation and implemented through registration agencies, of which Crossref is the largest. Crossref handles DOI registration for the scholarly publishing community — journals, books, conference proceedings, datasets, preprints, grey literature. When you see a string beginning 10. followed by a registrant code, a slash and a suffix, that is a DOI. The prefix identifies the registrant; the suffix is whatever the registrant chose, typically a journal code plus an article identifier. The whole string is registered with a resolver — doi.org — which performs the lookup and redirect.

This design means DOIs can be assigned to things that neither ISBN nor ISSN was built to handle: a single chapter, a dataset underlying a paper, a specific version of a preprint, a figure from within an article. Crossref has extended the system to cover grey literature and government reports. DataCite, another registration agency, specialises in DOIs for research datasets. The scope of "citable object" has expanded considerably beyond the journal article that was the original primary use case.

A DOI says nothing about who holds copyright in the object, nothing about its commercial status, and nothing about whether it is freely accessible. An article may be paywalled and still have a DOI; an open-access article still has one. The identifier is about persistence and citability, not about access rights or commerce.

§ 05Where They Overlap — and Where They Don't

A journal article, then, might legitimately carry: an ISSN (for the journal itself), a DOI (for the article as a citable, resolvable object), and — if the article is also published as a standalone offprint or the journal volume is later issued as a book — an ISBN too. These numbers do not duplicate each other; they answer different questions about the same intellectual content at different levels of granularity.

Understanding this is especially useful for researchers navigating databases, for publishers choosing which identifiers to register, and for anyone building a citation practice that needs to survive link rot. The ISSN tells you which publication. The ISBN tells you which product. The DOI tells you how to find and cite a specific object reliably.

None of them, it is worth repeating, is a copyright claim or a rights record. They are identification systems — infrastructure for knowing precisely what you are talking about, so that ordering, cataloguing and citing can work across institutions, platforms and decades. Rights travel separately: in licences, in registrations, in the work's own metadata. The identifier is simply the name that makes coordination possible.

APPENDIXAttached to the record

1. Late 1960sISBN developed from a British bookseller numbering scheme
2. Early 1970sISBN adopted as international standard
3. 2007ISBN migrated from 10 to 13 digits, aligning with EAN barcode family

Filed as general information, in a precise reference voice — not legal advice.