WorldCat and Union Catalogues

A union catalogue is one catalogue covering many libraries. Search it and you get a description of the work plus a list of the institutions that hold it, which turns two separate questions — does this exist, and where can I get it — into one.

That is a small idea with large consequences, and most of the machinery of modern librarianship is downstream of it. Shared cataloguing, interlibrary lending, consistent description across countries and the ability to ask what exists anywhere all depend on libraries agreeing to describe things the same way and put the results in one place.

The problem it solves

Cataloguing is expensive. Describing a book properly takes a trained person time, and the description produced in one library is close to identical to the one produced in another for the same book. Doing that work several thousand times over, once per library, is an obvious waste.

The second problem is discovery. A library’s own catalogue can only answer questions about its own shelves. For a reader whose subject is not held locally, the useful answer is a list of the places that do hold it.

Shared cataloguing solves both at once. One library creates the record, everyone else attaches their holding to it, and the accumulated file becomes both a cataloguing source and a directory of who has what.

How the shared record works

The mechanism is unglamorous and depends entirely on discipline.

A cataloguer searches the shared file first. If a record for that exact edition exists, they attach their library’s holding to it and stop. If it does not, they create one to the agreed standard, and it becomes available to everyone else.

Two conventions make this survive at scale. Records are written to a shared format — the MARC family, in most of the world — so that a record created in one system loads into another. And each record carries identifiers that allow the same work to be recognised across systems, which is the part that has changed most in the last decade and is taken up in Library APIs and Open Catalogue Data.

The underlying rules of description are the same ones an individual catalogue uses, and How Does a Library Catalogue Work? covers them.

WorldCat and OCLC

The largest union catalogue is WorldCat, run by OCLC, a non-profit library cooperative founded in 1967 as the Ohio College Library Center. Shared cataloguing on it began in 1971, and the membership grew from a state system to an international one.

Two things about it are worth understanding.

It is a cooperative, not a vendor in the ordinary sense. The records come from member libraries, and the value of the file is the accumulated work of the people who contributed to it. That is also the source of the periodic disputes about who owns the records and on what terms they can be reused.

And its holdings data — which library has which item — is the part that cannot be reproduced from anywhere else. Bibliographic descriptions exist in many places. A current, maintained map of who holds what does not.

National and regional union catalogues

WorldCat is not the only model, and in much of the world it is not the primary one.

National libraries frequently run a national union catalogue covering libraries in their own country, which is the file that matters for domestic interlibrary lending and for legal deposit. Regional consortia run shared catalogues covering a group of institutions, often with a shared lending agreement attached, so that a reader at one member library can request from any of them. Subject-based union catalogues exist for particular fields, especially where the material is specialised and the holding libraries are few.

Most libraries participate in more than one, at different scales, for different purposes. The differences between the institutions involved are the subject of Types of Libraries.

Interlibrary lending, which is the point

The practical payoff is that a reader can request something their library does not hold.

The sequence is the same everywhere: the item is located in the union catalogue, a request goes to a holding library, and the item is sent as a physical loan or, for an article or a chapter, as a scan supplied within copyright limits. The reader collects it from their own library.

This is the service that makes small libraries usable for serious research, and it is invisible to most people who could benefit from it. It is also the reason accurate holdings data is not an administrative nicety: a request sent to a library that no longer holds the item wastes everybody’s time.

What is changing

Three pressures are reshaping union catalogues at once.

The record format is being replaced. MARC was designed in the 1960s for printed cards on magnetic tape, and successors based on linked data are being built to let bibliographic descriptions work as web resources rather than as files exchanged between library systems.

Identifiers are doing more of the work. Matching records by comparing text is unreliable; matching them by shared identifiers for works, editions, people and organisations is not, and this is where library data increasingly connects to the open web.

And the cooperative model itself is under argument, over cost, over who may reuse the records, and over whether an open alternative could hold the same ground. Digital Library covers the broader version of that question.

Sources and further reading

OCLC documents its own history and the structure of the cooperative, and WorldCat is searchable without a subscription. The Library of Congress publishes the MARC standards and the specifications for their successors. Wikipedia’s article on union catalogues gives the general background with onward references.