140 Personal Content Experience: Managing Digital Life in the Mobile Age desired (this technique is one form of “query by example” technique), or by translating the extracted features into keywords. In this case, all techniques available for textual searching can be subsequently applied. In many cases, CBR poses further challenges to user interface design (section 6.5). What is especially interesting in CBR is the way in which the search criteria are provided. This may be relatively straight- forward with visual information, but how is one to provide an example of a specifi c kind of heart rate curve? Or a saved game, for that matter? Yet another aspect related to searching is the searching range. This is especially relevant to mobile use with distributed content; specifying whether the content is searched for in the device only, from the whole personal device ecosystem, or from the Internet, heavily affects the performance and thus perceived effi ciency of the whole system. Furthermore, performing a search on a peer-to-peer network, where the results indicate currently available content, is inherently different compared to searches on an indexed database (such as current web searches), which may contain outdated information on content availability. 5.3 Different Types of Metadata Prior to diving into the mysteries of metadata management, some dis- cussion on metadata classifi cation is in order. Next, we will introduce our way of classifying metadata. Metadata is not just something that is glued onto the data it is describing, but is inherent, built in every content object, system component, and piece of software in a computer device. Metadata related to a content object is as much a living thing as the object itself. Usually we regard metadata as simple attributes or tags that describe the content object to some extent. For example, all fi les have a creation date and a name attribute, providing information on when the fi le was created and how it can be referred to in the system, respectively. Other examples of tag-like metadata are album and artist information in MP3 fi les, or the aspect ratio of a feature fi lm on a DVD. What else would one need to gain a suffi cient access to and control over their personal content? We have expanded the concept of metadata far beyond mere tags, as our four categories of metadata illustrate: