How Does a Library Catalogue Work?

A library catalogue is not a list of books. It is a database of records, one for each item, each written to a shared standard so that a search can match on any part of it and so that libraries can exchange descriptions instead of writing them all from scratch. When you type a title into the search box, you are searching those records, not the books.

Everything else follows from that. The catalogue tells you what the library has and what each item is; classification tells you where it physically stands; identifiers tell the world which exact edition is being discussed.

What is in a record

A catalogue record describes one item in a structured way, with each piece of information in its own field: title, statement of responsibility, edition, publisher, place and date, physical description, language, series, notes, subject headings, and identifiers such as the ISBN. Attached to it are holdings, which say how many copies the library has, where each one is, and whether it is on loan.

Structure is what makes the record useful. Because the title sits in a title field and the date in a date field, the catalogue can offer search by author, limit by year, sort by date and distinguish a book about an author from a book by them. A flat block of text could not do any of that.

Records are exchanged in defined formats. For most of the last half-century that meant MARC, the family of machine-readable cataloguing formats, and the field is now working towards linked-data models that express the same information as relationships rather than as fields. The description a small library uses very often originated at a national library, which is the practical value of a shared standard.

Authority control, the part nobody sees

The hardest problem in a catalogue is sameness. One author writes under a maiden name, a married name, initials and a pseudonym; two different people share a name; a place has been spelled three ways; a subject was called one thing in 1970 and another now.

Authority control solves this by maintaining a single authorised form for each name, place, work and subject, with the variants recorded as pointers to it. That is why searching for an author usually returns everything by them regardless of how a particular title page printed the name, and why a subject search can pull in material that never uses your exact words.

It is slow, expensive, invisible work, and it is the main reason a catalogue can outperform a keyword search over the same collection. It is also inherited wholesale by digital collections, as What Is a Digital Library? describes.

Classification puts it on a shelf

Classification is a separate job from description. It assigns each item a place in a subject scheme, so that everything on a topic stands together and browsing works. Two schemes dominate in English-language libraries.

The Dewey Decimal Classification divides knowledge into ten broad classes numbered 000 to 900, each subdivided decimally, so a longer number means a narrower subject. It is compact and widely used in public and school libraries, where the collection is broad and shallow.

The Library of Congress Classification uses letters for its main classes with numbers beneath them. It expands more comfortably into very large and specialised collections, which is why academic and research libraries tend to use it.

Neither scheme is neutral. Both were built at a particular time from a particular view of how knowledge divides, and both have been revised repeatedly where the original arrangement misplaced or marginalised subjects. Many libraries also depart from the scheme deliberately, shelving children’s material, fiction or local collections by their own logic because that is what their readers actually browse.

The call number is the result: a class number plus letters standing for author or title, and sometimes a volume, copy or location code. It gives one item one address, which is what makes retrieval from a large collection possible at all.

Identifiers name the exact thing

An ISBN identifies one edition of one publication in one format from one publisher. The hardback, the paperback and the ebook of the same title each have their own, and a reissue with different pagination will too. That specificity is the point: it lets a library order, receive, lend and report on a precise object without arguing about titles.

Other identifiers do the same job elsewhere: ISSN for serials, DOI for articles and datasets, and various persistent identifiers for authors and works. Together they are the plumbing that lets separate systems talk about the same item.

Using it well

Most public catalogues now present a single search box and then hide considerable power behind it. A few habits make a large difference.

• Search by field when you know something exact. Author or title searching cuts through the noise a keyword search generates.
• Use the facets. Format, language, date, location and availability filters on the results page are usually faster than reformulating the query.
• Follow the subject headings. Opening a record you like and clicking its subject terms is the quickest route to comparable material, because it uses a librarian’s judgement rather than your vocabulary.
• Read the holdings. The record says which branch has it, whether it circulates and whether it is reference only. A great many failed visits are explained there.
• Ask for what the library does not hold. Interlibrary loan and consortium requests bring in material from other collections, and the catalogue often searches those partners alongside its own.

Why this still matters

A catalogue is slower to learn than a search engine and answers a different question. It tells you what exists in a defined collection, in which edition, and where the copy is, and every part of that claim can be checked. That is exactly the guarantee a generated summary cannot offer, which is one of the threads in Will AI Change Libraries Forever?.

Understanding the machinery also makes the building easier to use: once the call number and the subject heading make sense, the shelves stop being a wall of spines. The Modern Library picks up what happens once you can find things.