Chapter 8: Timeshifting Life 335 maintaining the information. A more realistic approach is to build partial ontologies in the most important domains, and start to build the semantics based on these. Another crucial issue is metadata interoperability. In case the exist- ing numerous metadata formats continue to be, more or less, incom- patible in terms of both defi nitions and semantics, the interoperability between applications and cross-media content remains a goal out of reach. An effi cient mechanism for conversions between formats, as well as standardized ontologies, is an essential prerequisite for meta- data to really start gluing the content objects together. The quality of metadata is another signifi cant factor that is refl ected both in technological and user-related categories. Fragmented meta- data is a particularly challenging issue. When some metadata exists, but is scattered sparsely across personal content, it provides too little information for solid automation. In a similar manner, cross-media, cross-application, and cross-manufacturer metadata formats and stan- dards (that is, universal metadata management) are a challenge, since dozens of players from various domains are involved. Context recognition and processing technologies may develop more slowly than anticipated, which will lead to starvation of context-hungry applications. This is a signifi cant threat, since without fi le context (section 4.7) many features of smart metadata, relations, and social aspects become harder to implement. Problems may also arise from unexpected directions. For instance, the increasing growth rates of personal content, and the expected growth of related metadata, require an increasing amount of mass storage and memory. We are used to thinking that the price of storage (bits per dollar) will keep decreasing but, for instance, a single fi re in a chip foundry can signifi cantly disturb chip production, drive fl ash memory prices up, and thus slow device sales to a (temporary) halt. 8.4.2 Human-Related Issues One of the greatest challenges related to users and metadata is the fear of compromised privacy. Once the amount of personal content, together with potentially extremely intimate metadata, increases and spreads across devices, it will probably raise fears and uncertainty: how well will the information stay private? This is a justi- fi ed concern, and should be addressed both in system design (that is providing suffi cient methods for fi ltering metadata, encrypting it, and so forth) and user interface design (that is, making it explicit that the metadata and personal information is indeed kept safe and sound).