Both a classified content and unstructured data view are valid and necessary. Organizational principles are the domain of librarians creating classifications and ontologies including the semantic web versus machine learning approaches to self-organize content. However, with large quantities of information, users are not just unwilling to classify, but are in fact unable to do it. Special skills are required to construct useful classifications. The first impulse we and others have felt is to just wish away the usefulness of metadata and hierarchies. But full-text search is not enough; in our experience, many items require some other attributes to be found
