Subject Classifications

The top two levels of the American Library of Congress subject classification is provided by default with a new EPrints archive.

Classification File

The classification is archive-specific and is stored in the ~/eprints/archives/cfg/subjects data file. It has the following format

subjects:Library of Congress Subject Areas:ROOT:0
A:A General Works:subjects:0
AC:AC Collections. Series. Collected works:A:1
AI:AI Indexes (General):A:1
AM:AM Museums (General). Collectors and collecting:A:1
AS:AS Academies and learned societies (General):A:1
AZ:AZ History of Scholarship The Humanities:A:1
B:B Philosophy. Psychology. Religion:subjects:0
C:C Auxiliary Sciences of History:subjects:0

Coding a Classification

Internally, the classification is a set of keys (e.g. A, QA76 etc)
• which have human readable labels (e.g. "A General Works", "QA76 Computer Software")
• each of which has one (or more) parents.
• There is a special 'top level' entry called ROOT which has no parent.
• Each classification entry is marked to indicate whether it can have documents attached to it, or whether it is just a placeholder.

Altering the Classification

It is easy to alter the available subjects, either
• by filling in missing detail in particular parts of the hierarchy,
• by pruning unrequired parts of the hierarchy
• or by completely replacing the subject hierarchy with one that is more relevant

Adding new keys has no consequences for the running archive,
• but deleting them could possibly leave some eprints using a now-non-existent key.

After any changes to the subject file, it is necessary to run import_subjects

Grafting on a Separate Classification

It is possible to run two independent classification schemes by joining them at the ROOT.
• (It may be useful to have two or more different ways of classifying an eprint,
• for example, by subject and by institutional research group)
• Each subtree (here departments and subjects) can be specified as the type of a metadata field in the database.