When most people think about library classification, names like the Dewey Decimal Classification or Colon Classification come to mind. These are systems built to arrange books neatly on shelves. But there is a lesser-known scheme that was never really meant for shelving at all. The Broad System of Ordering (BSO) was created to solve a very different problem: how do you get two information systems that “speak” completely different classification languages to understand each other? This question became urgent in the 1970s, and the answer that emerged was one of the most ambitious experiments in knowledge organization of its time.
Table of Contents
What is the Broad System of Ordering?
The Broad System of Ordering is a general, faceted classification scheme designed for the exchange and switching of information between systems. It was commissioned by UNESCO and developed by the International Federation for Documentation (FID), with work beginning around 1971 and a major schedule appearing later in the decade. The initial drafting was led by Eric J. Coates along with G.A. Lloyd, and the third revised edition was published in 1978 by the FID.
The key word in the name is “broad.” Unlike schemes that try to pinpoint a narrow subject down to the smallest detail, BSO deliberately stays at a higher, more general level. It was built to meet the requirements of the UNISIST programme, an international effort to connect scientific and technical information systems across the world. The goal was not to replace existing classification schemes but to act as a bridge between them.
This makes BSO a “switching language.” Think of it this way: if one library indexes its records using the Universal Decimal Classification and another uses a specialised thesaurus, the two cannot directly read each other’s subject codes. BSO sits in the middle as a common reference point that both systems can map to, allowing information to flow from one to the other.
BSO structure and notation
BSO organises knowledge into a hierarchical arrangement of subject fields. In its developed form, the scheme covers several thousand subject fields grouped under a manageable number of main classes, moving from the formal and physical sciences through the life sciences and on to the humanities and social sciences. The structure is intended to give a complete picture of human knowledge while staying flexible enough to handle subjects that cross disciplinary boundaries.
A faceted approach
BSO is a faceted classification. This means that instead of listing every possible subject in advance, it breaks subjects down into component parts, or facets, that can be combined as needed. A facet is simply a single characteristic by which a subject can be divided, such as the process involved, the object being studied, or the place and time. This faceted design is one of the features that links BSO to the analytico-synthetic tradition associated with Ranganathan’s facet analysis, where a subject is first analysed into its parts and then synthesised into a class number.
Interestingly, BSO does not always follow the same facet sequence as Colon Classification. In some areas it places the parts of a subject and the objects of study on the same level, which changes how composite subjects are built compared with Ranganathan’s method. This shows that BSO borrowed faceted principles but adapted them to its own switching purpose rather than copying any single scheme.
Notation
The notation of BSO is mostly numeric. Subject classes are represented by short numeric codes, supplemented in certain cases by letters to indicate common facets or aspects that apply across many subjects. The designers kept the notation deliberately concise and easy to remember. The reason was practical. Because BSO was meant to map between other systems rather than serve as a detailed shelving tool, a short and simple code was more useful than a long and highly specific one. A notation that is hierarchical also helps because broader and narrower subjects can be recognised from the structure of the code itself.
BSO’s role in information systems
To understand why BSO looks the way it does, it helps to recall the environment of the 1970s. Information was growing rapidly, and many countries and organisations were building their own databases and indexing systems. Each used its own subject vocabulary. There was a real fear that this would create islands of information that could not talk to one another.
BSO was designed as the connecting tissue. As a switching language for information exchange, it allowed records indexed under one scheme to be translated, or switched, into another. A database could tag its contents with BSO codes, and any other system that also understood BSO could then locate relevant material without needing to learn the first system’s native classification.
Where BSO was actually used
Because of this purpose, BSO found its home in international information networks rather than in ordinary lending libraries. It was associated with UNESCO’s information programmes and was explored by large cross-national systems as a way of exchanging bibliographic information across institutional and national boundaries. Major international databases in fields such as nuclear information and agricultural science looked at BSO as a possible means of improving retrieval across systems, even though full adoption remained limited.
In the library world more narrowly, BSO had very little uptake. A few libraries experimented with it, but it never became a mainstream shelving scheme. This was never a failure on its own terms, because shelving was never its goal. The scheme is sometimes described as a “roof classification,” sitting above other systems rather than competing with them at ground level.
Challenges and impact
Despite its inventive design, BSO never achieved wide adoption. Several factors explain this.
Why adoption stayed limited
First, the idea of a universal switching language is technically demanding. Mapping one classification scheme to another is difficult because schemes divide knowledge differently, use different levels of detail, and treat interdisciplinary subjects in their own ways. A broad intermediate scheme inevitably loses some of the fine detail that specialist systems rely on, which made some institutions reluctant to route their data through it.
Second, BSO arrived at a time when computing was beginning to change the rules. As databases became more powerful, many organisations preferred to invest in their own indexing and in direct mapping or thesauri rather than maintain a separate shared scheme. The administrative effort of keeping a switching language up to date across many participating systems was considerable, and responsibility for the scheme passed through bodies such as the FID/BSO Panel over the years.
Third, established schemes had momentum. Libraries and information centres had already invested heavily in systems like DDC and UDC. Switching to or even alongside a new scheme requires training, retrospective conversion, and ongoing maintenance. For many, the practical benefits of BSO did not clearly outweigh these costs.
The lasting impact
Even though BSO did not become a standard, its influence should not be dismissed. It was one of the clearest early attempts to take the problem of interoperability seriously, decades before the term became common in digital information work. The challenge BSO tried to solve, namely getting different systems to share and translate subject information, is exactly the challenge that modern efforts in linked data, metadata crosswalks, and knowledge organisation systems continue to face today.
BSO also stands as a useful case study in classification theory. It demonstrates how a faceted, general scheme can be built for a purpose other than shelving, and it highlights the trade-off between breadth and detail that every classification designer must confront. For students of information processing and retrieval, understanding BSO means understanding why interoperability is hard and why the dream of a single connecting language is both attractive and difficult to realise.
How BSO compares with traditional schemes
It is worth drawing a clear line between BSO and the classification schemes most students study first. The Dewey Decimal Classification and the Library of Congress Classification are primarily designed for arranging documents in physical space. Colon Classification is faceted and analytico-synthetic but is still aimed at producing specific class numbers for individual documents. BSO shares the faceted method but applies it to a different end, producing broad codes meant for translation between systems rather than for precise shelf location.
This difference in purpose explains almost everything about how BSO looks. Its notation is short because long notation would defeat the goal of easy mapping. Its classes are broad because fine detail varies too much between systems to translate reliably. Its facets are flexible because interdisciplinary material is exactly the kind of content that switching between systems must handle. Seen this way, BSO is less a rival to Dewey or Colon and more a tool meant to work behind the scenes, connecting the systems that use those schemes.
What do you think? If BSO was built to solve the problem of systems that cannot understand each other, do you think a single switching language can ever fully capture the detail that specialist systems need? And given the rise of digital metadata standards, is the idea behind BSO outdated, or more relevant than ever?
Leave a Reply