If you'll pardon the militaristic metaphor, we can observe that a few Semantic Web SeedApplications exhibit a common pattern: opposing use cases share common requirements, data structures, and descriptive needs.
The militaristic metaphor is perhaps appropriate: the Web provides a pluralistic architecture for mediating disagreements, dispute. There is only one Web, and it serves a variety of sometimes competing purposes and communities.
A (can we drop the 'Semantic' qualifier) Web principle: unexpected re-use is inevitable. We can't anticipate all the uses to which our descriptions and identification strategies on the Web we be put.
The impact of the Web cuts both ways. When we richly describe Webcontent to protect the rights of copyright holders, we provide metadata which file-sharing communities might use (knowingly or otherwise) to discover, index and redistribute that content. When we richly describe Web content to help protect children from stumbling across pornography, we simultaneously provide technology to support its discovery and dissemination in general.
We can explore such tensions by looking at a couple of newsworthy Internet issues: concerns around content filtering vs privacy and around file sharing vs rights management. Perhaps there are others, but these should be plenty to be getting along with. (Over)generalisation can come later.
The content filtering DoubleEdgedWebSword: many people and organizations are concerned about the widespread and easy availability of 'adult' (pornographic etc) content in the Web. This was the motivation for W3C's work on the Platform for Internet Content Selection (PICS), the precursor technology to RDF, aka "PICS-NG". ContentSelection4SemanticWeb provides further historical background. PICS provided a way to label Web content as 'adult', 'violent' etc. Such labels can (but are they?) be used to support searching of Web content based on exactly those properties, ironically making it more readily accessible to those with unfiltered access to the Web.
The file-sharing double-edged sword: many people and organisations are concerned about the widespread and easy availability of copyrighted content (MP3 encodings of songs, etc) in the Web. This has motivated much work on rights description vocabularies. Some (such as CreativeCommons) motivated by a desire to allow 'some rights restricted' principled sharing of content (eg. personal photo archives for non-commercial use). Others, eg. the INDECS initiative, led from the perspective of rights-holders who want to express precise and unambiguous rules for the re-use and re-combination of media content, with particular concern for tracking the contributors to each sub-component of some content. Such descriptions are rich and sophisticated, because there is an economic driver: we want to know who played drums on which track of an album, so they can get paid when a recording is played on the radio. But such rich descriptions, if widely available and associated with identifying information for copies of the content, may make it easier for file-shares to find and share the content, regardless of rights and restrictions. In particular, the establishment of widespread conventions for identifying artists and their works (eg. album and movie titles) is likely to benefit both rights holders and file-sharers...
where is this going? some example markup would be good... --DanBri