I’ve long pondered why the records world doesn’t use or understand thesauruses properly. Twenty years ago a library user searching for automobiles would be told – by the software – that the preferred term was cars, and shown the results for a search on cars. What’s more, the same functionality provided guidance through the maze of organizational history: earlier and later names, splits and mergers. A thesaurus is one of a family of terms including ontology, taxonomy and classification. But if it lacks semantic guidance, it’s not a thesaurus, it’s a file plan, or, in specialized form, just a disposal schedule. In records, we don’t even have this semantic functionality in most of our software systems.
Even thesauruses, as defined by the ISO standards, are now inadequate. Most organizations have clusters of terms which demand several equivalent synonyms. It is a distortion to say that one of these must be the preferred term, because it depends, for example, if you are in the legal, technical or public domain. A century ago, faceted classifications using index cards were common in research organizations throughout Europe, and we are only now able once again to provide that flexibility to our users, through tools such as folksonomies.
