| |
  |
| Is the end near for the file/folder metaphor? |
|
 |
|
posted by Editor
on Monday February 27, @08:43PM
|
|
|
|
|
The hierarchical directory (i.e. "file folder") method of representing data is so entrenched that it is hard to imagine this approach won't be used forever. But while the file folder model been very successful in the context of single user machines and small workgroups of users, it may not be suited for effectively managing data on the scale of global networks like the web. In this environment, the concept of a "file system" may evolve into more scalable data management models based on relational or temporal structures, in which users tag data with a variety of attributes reflecting its context and meaning, rather than referencing it in terms of a unique file name. As data becomes free, both in terms of its control and the cost of its storage, value will increasingly shift to the control of such metadata (i.e. "data about data"), which can be shared through peer-to-peer searches and explored with visual user interfaces. However, it is debated whether global metadata can be reliably assigned in a public network, or can only be extracted from implicit factors. Some experts are betting on the former with initiatives such as the semantic web, but critics charge that any labelling system will be at risk of corruption by off-topic tags (e.g. spam) or insincere tags (e.g. trolls). By contrast, leading-edge search interfaces such as Google have been extraordinarily effective at eliciting the "meaning" of data by analyzing its text and the web links connecting it, rather than depending on users to categorize the data. Still, the concept of collaborative categorization continues to generate interest, as shown by the emergence of folksonomies and the popularity of "tagging"-oriented sites such as del.icio.us and flickr. One possible defense against the corruption of shared metadata would be to validate it with distributed trust metrics. The integration of such loosely consistent mechanisms directly into applications would allow them to process data in a way that accounts for varying credibility.
|
|
|
|
< Will IBM's Cell Processor Launch An Innovation Boom?
| Skitter Represents Web Performance Visually >
| |
|
