One way in, three ways out

Lessons on sorting and finding from library science

(originally posted 11.16.17)

Information architecture, the logical organization of information, is usually associated with classification of digital information. However the most famous information architect may be Melvil Dewey, who published the Dewey Decimal Classification system in 1876. The Dewey Decimal system works on the premise that by using a single sorting characteristic, the topic of the book, one knows exactly where to sort any book, and exactly how to find it. While newer book classification systems have since been developed, the Dewey Decimal system remains a leading library information classification system around the world, due to its simple and comprehensible information hierarchy[i].

To sort any book using the Dewey Decimal system, the curator determines which of the ten 100-level categories the book belongs to, then which of the ten 10-level subcategories, and finally which of the ten 1-level sub-subcategories, resulting in a 3-digit number reflecting the topic to which the book belongs (with decimal places to further narrow the topic)[ii]. To find a book describing how to perform calligraphy, first locate the shelves holding the 700 category (“The Arts”), then the 740 subcategory (“Drawing & Decorative Arts”), and finally 745 (“Decorative Art and Design”). If this subsection holds numerous books, material on calligraphy in particular can be found in 745.6 (“Calligraphy, Illumination, and Heraldic Design”). While the system requires learning the basic categories and where to look up specific subcategories, it can be used by both curators and readers to sort and find information to a very fine degree of specificity.

And it is a fine system if one visits the library to find a book on a given topic. But what if the goal is slightly different, such as to find a book by a particular author, or a specific book by title? An index is necessary when the reader wishes to search for a book by a different characteristic than which it was sorted. The card catalog system long served as this index, providing a way to find any book not only by, but also by author or title[iii]. Each book had 3 separate cards, and every library had 3 sets of card catalogs, sorting the information by topic, author, and or title. Of course these nostalgia-inducing beautiful rows of wooden cases have largely been replaced by computerized book-finding systems, but the point illustrated remains the same: while a successful classification system relies on sorting information using a single method, there often needs to be more than one way to get the information back out. The classification system works by defining what the top level of information hierarchy is, and what information is considered to be of lesser significance. But the people using the system may hold a different characteristic to be the most important, and must be able to sort through the information based on a shuffled set of hierarchical rules.

In our digital age, the amount of information in a system or website can be immense. But sorting principles can be applied from the system developed nearly 150 years ago, which remains capable of organizing every book that has ever been published. As the curator, the information architect must determine the most logical characteristic by which to arrange the information by understanding what the users of the system are most likely to be seeking. This characteristic becomes the top-level of the information hierarchy, and with additional subcategory characteristics, will form a set of guidelines used to place every piece of information in the system. The display of information will be consistent and any piece findable through searches by the defining characteristic.

But as with readers in a library, users may have different goals than the information architect originally imagined. Like the indexes in the card catalog, digital information should be virtually indexed by any defining characteristics other than the one by which they were sorted. In most systems, there should be only one way for information to be sorted in to a collection, but several ways by which it can be searched out. Although the format and the amount of information has changed over 150 years, these basic tenets of the Dewey Decimal System remain relevant to today’s information architecture.

[i] Dewey Decimal Classification Summaries:

[ii] Dewey Decimal vs Library of Congress: