Subject metadata conveys what a resource is about (see my previous entry on isness and aboutness). The two types of formal subject metadata in use are subject headings and descriptors.
By formal subject metadata, I mean metadata that is assigned by trained metadata specialists using a controlled vocabulary, as opposed to tags supplied by “the cloud,” that is, people using and visiting websites that allow users to assign tags to digital objects.
First, let’s look at subject headings:
Subject headings are comprehensive in coverage; they describe the content of an entire digital object or information resource. In practice, subject headings are assigned, for example, for an entire book.
Subject headings are designed to be displayed in an alphanumerically sorted browse display. In the past, librarians have used the term “pre-coordinated” to describe subject headings of this type. It basically means that the subject heading’s elements are arranged in a prescribed order, like this:
Hydrology– Montana–Cascade County–Maps.
This arrangement has the advantage of enabling the user to truncate the search. A searcher doesn’t have to know the whole string in advance. In this case, a searcher could only search “hydrology” and the search results would be a browse index that starts at hydrology. Then the searcher could scan the display looking for the desired information.
Generally, only a few subject headings (1-6) are assigned per book (or other resource).
Next, let’s look at descriptors:
Descriptors tend to be used for shorter works such as journal articles.
Each descriptor describes one of the concepts that the article (or other work) talks about. In other words, descriptors don’t necessarily reflect the coverage of the entire document they describe, just a part of it.
Unlike subject headings, descriptors aren’t designed to create alphabetical browse displays. Instead, they are designed to accommodate Boolean searching. Here’s an example of a search:
FIND hydrology AND Montana AND maps
This type of search is called post-coordinate, because the searcher (not the indexer) puts the terms together in the search. Descriptors aren’t free-form like tags in folksonomies. Instead, like subject headings, they come from a controlled vocabulary. But because descriptors describe content at the concept level, there tends to be more of them used to describe a document — much more than subject headings. It’s not uncommon for there to be twenty or more descriptors for one document.
The irony is that she shorter a work is, the more descriptors it tends to have, and the longer a work is, the fewer subject headings it tends to have.
These two examples illustrate this unevenness:
Article: Vestibular pneumolabyrinth: why assessment with temporal bone computed tomography utilizing dynamic focal spot mode is important for the diagnosis.
Book: The History of Portugal.
So depending on the subject metadata needed for a document collection, both subject headings can be useful depending on the content that is indexed and the needs and abilities of the searchers.