142 Personal Content Experience: Managing Digital Life in the Mobile Age ID3 (key, value) pair can be expressed as “artist = Schbert”. The key, here “artist”, defi nes meaning for the value, here “Schbert”, saying that Schbert is the creator of the song. Recently, another way of using tags without clearly defi ned seman- tics has arisen. These so-called folksonomies, are tags without a key or rules on how to use them (section 4.1). Schbert’s song “But not today,” could be expressed simply by listing of plain tags, i.e., just values without keys: Schbert, Pest of Schbert, 2006. There is no way to understand what each tag means but usually music listeners can quickly decipher that “But not today” is a song by Schbert and it is in the album Pest of Schbert released in 2006. Folksonomies are extensively 11 12 used in services like YouTube or Flickr, where consumers create and tag the content themselves. Users see what tags others use and start using the same tags themselves to describe the same properties. Usually, these services also automatically link content with the same tags to each other, thus creating groups of similar content. Folksonomies are easy and fl exible to use since the user can choose whatever tags he or she likes or thinks describes the content best. They are useful for humans but extremely diffi cult for computers, since there is no predefi ned semantics behind them. We believe that for effi cient metadata management both approaches, i.e., clearly defi ned seman- tics and freedom to create your own tags, are crucial and all metadata management frameworks must support them. 5.3.2 Context Capture You do not usually think much about metadata while taking your vaca- tion photos, even though metadata plays such an important role. At the click of a button you freeze the image in front of the camera, but the visual image contains only a tiny part of the prevailing reality. There may be friends hanging around, yet not all may be visible in the photo. Furthermore, the image does not usually identify the location, or the name of the mountain you use as a backdrop. All that information, generally called context information, is usually lost for good. While it is important to capture the images or sounds, for the purpose of re- living the experience it is equally vital to capture the whole situation, too. As pointed out in section 4.7, there are a number of potential context information sources, some of which are easier to obtain than others (Figure 5-1). Our framework stores available contextual 11 http://www.youtube.com/ 12 http://www.fl ickr.com/