MCMXLIX / 1949
The Outline of History
USBNUAZJA136WFYXF
WSBNWC17225YANQAM
The Universal
Standard Book Number
A deterministic, registry-free identifier for books — computed from the title page alone. No central authority. No fees. No gatekeepers.
UAZJA136WFYXF
WC17225YANQAM
Try
The ISBN system, standardised in 1970, has given every new book a number. The roughly sixty million editions that came before have no such identifier, and the only alternatives — OCLC control numbers, Library of Congress control numbers, Open Library identifiers — all require a registry, a membership, or prior cataloguing by someone else. A used-book dealer holding a 1923 volume that no institution has previously catalogued has no standard identifier to point to at all.
USBN is the alternative: compute the identifier from the book itself. Hash the title, author, and year — as printed on the title page — into a thirteen-character string, using an algorithm anyone can run on a laptop or in a library terminal. The same book, examined by two strangers on opposite sides of the world, yields the same USBN every time. No network. No registration. No fees.
How a USBN is computed
-
Read the title page
Transcribe the title, author, and publication year as printed. Diacritics, case, and extra whitespace are normalised away so two catalogers in different locales produce byte-identical input.
-
Hash with BLAKE2s
The normalised string is hashed with BLAKE2s to produce a 64-bit digest; the top 60 bits are taken as the book's fingerprint.
2⁶⁰ ≈ 1.15 × 10¹⁸. -
Encode in Base32
The 60-bit fingerprint is written in Crockford Base32 — twelve characters — and prefixed with
U. The result isU + 12 chars = 13 chars, exactly the length of an ISBN-13.
Some books with their USBNs
MCMLIX / 1959
The Elements of Style
William Strunk Jr. and E. B. White
USBNU4TMJP8GE1DSF
WSBNW4NPQT7637D53
MCMXVI / 1916
Über die Relativitätstheorie
Albert Einstein
USBNURAYHF9EDXKGQ
WSBNW718QV0NXA405
MCMLXIV / 1964
College Calculus with Analytic Geometry
Murray H. Protter
USBNUGM4Y9KZVGYH7
WSBNWDYNK8KP7FHSG
Notice that different printings of the same work share a
WSBN — a companion work-level identifier —
while each edition still has its own unique USBN.
What USBN is good for
- Used-book dealers get a common identifier across platforms without coordination or API integration — every seller computes the same number from the same edition.
- Libraries can catalogue pre-ISBN holdings and link them across union catalogues without waiting for OCLC membership or record assignment.
- Digital humanities projects can cite pre-1970 texts with stable, reproducible identifiers instead of fragile URLs and drifting database keys.
- Personal libraries get a portable, tool-agnostic identifier that survives the migration from LibraryThing to Goodreads to whatever comes next.
- Citations and footnotes can include a USBN alongside the prose reference, enabling future reconciliation with any bibliographic database.
The paper
Technical paper · April 2026 · English
USBN: A Deterministic, Registry-Free Identifier for Pre-ISBN Books
Lennart Lopin · Euler's Identity LLC
Submitted to The Code4Lib Journal
The formal presentation of USBN v1.0. Introduces the algorithm and the canonical normalisation pipeline; analyses collision probability against the estimated sixty-million-edition pre-ISBN corpus; documents the four defects in the original draft that motivated the v1.0 rewrite; argues for thirteen characters on the grounds of ISBN drop-in compatibility; includes complete reference implementations and a full test vector table.