Beautiful Data

Directories: Priceless

Think about a library. Think of the library's floors, hallways, and shelves as silos of information. Valuable information is tucked away, waiting to be discovered. No one roams the halls when looking for a specific book. Instead, one uses the card catalog—cross-referenced on subject, title, and author—to facilitate discovery of relevant documents.

Let's say that directories, indices, and catalogs are all basically the same thing: a thing used to locate other things. Some examples of locators include the card catalog at the library, phone directories, Google, eBay, and so on. In each case, the directories are equivalent to locator services: they return reference information (pointers) after being provided one or more search terms. At the library, the card catalog is a special-purpose directory used to enable efficient enterprise discovery, providing the user a specific pointer to a document (e.g., using the Dewey Decimal system). After the user is provided a pointer, the activity becomes "federated fetch." Note the difference between federated search (not useful) and federated fetch (useful).

A Google search does not scour the planet for the results; rather, a specialty directory created by Google is searched and the results, pointers to the real documents (e.g., URLs) are returned to the inquirer.

It would follow that the only scalable solution to enterprise-wide discoverability involves the use of directories. No surprise there; a special-purpose directory is ...