Chapter 4: Metadata Magic 117 That would not matter, but what really brings her down is that she can’t use keywords in the upload client. She has to fi rst select the images by keywords in the browser, export them as a group, and only then start the upload process. It would be better for her way of working to just start the upload client and select the images from there. Cathy and computers are not a good match, she feels. But there’s no going back. Steve – help! So, universal content metadata could enhance moving content between applications. Briefl y, it is essential to provide the semantics of each metadata attribute to allow application writers to use the metadata, and to allow the metadata being converted from a format to another. One desirable aspect of universality is extensibility. The metadata format should not be carved in stone, as new content types, and new uses for old ones keep coming up. Therefore, the formats should allow adding new fi elds, new data types to existing fi elds, and perhaps redi- rection or even redefi nition of existing fi elds. This may be an intractable problem; however, some existing formats (such as MARC) successfully incorporate extensions in their basic mechanisms. Our approaches to extensibility are discussed in Chapter 5. Last and defi nitely not least, there is one more challenge that is not technical, but human. The challenge we are referring to is abuse. Unfortunately it seems that regardless of planning, protection, and technical solutions, you cannot underestimate the amount of work some people are willing to invest in order to create something that is harmful. There are various reasons for generating low-quality, falsifi ed, and wrong metadata, such as ignorance, mischievousness, and greed. Mischievousness is generally about producing metadata that may be insulting, absurd, and irrelevant intentionally. This phenomenon is actually ancient and part of human nature; small numbers of people tends to get pleasure from ruining other people’s work. Also, this is something that plagues many open and shared Web services, such as Wikipedia. Ignorance is related to the fact that people are not aware of the possibilities of the metadata and use it carelessly. In contrast, greed is an opposite case: some parties who are abusing metadata are well aware of its power. A good example of the effect of greed is that initially the Internet search engines used metadata that was embedded in the Web pages. Soon some (adult) content and service providers noticed this and started the vast abuse of the keywords in order to get more visibility to their product in search results. As a result of this abuse, the