Electronic archives may contain very large amounts of data. In such cases, the individual datasets are stored in accordance with certain indexing keys, which is where a distinction must be made between business application index data and archive system index data. If unsuitable classification criteria are used, it may be difficult or impossible to search for archived documents. It may also be impossible to clearly determine the semantics of documents. Furthermore, an unsuitable or limited selection of classification criteria could thwart the objectives of preservation (e.g. the ability to provide evidence to third parties).