RSS is extensible. Ben is showing RSS that contains contributor information (using Friend of a Friend data inside Dublin Core contributor tags), trackback data, and annotation references. All of this is added automatically by the system.

ThreadsML is based on RSS 1.0. It uses an Apache module called mod_threads and Dublin Core. Using ThreadsML, conversation happening in blogs, mailing lists, and IRC could all be merged together to form a threaded stream of the conversation.

The major problem with this is subject.

• Subject lines are rarely useful
• One old solution is hierarchal ontologies
  • Easy to use but...
  • It only works one way.
  • Hierarchical ontologies are cultural specific.

One answer might be ENT or Easy News Topics for RSS 2.0. ENT allows ontologies to be created on the fly. This takes away the cultural specificity and allows each person to define and publish their own classification systems. (Here are notes about a breakfast conversation on ontologies that I didn't take part in.)

To make this useful, there needs to be a way to link ontologies. There's a big question about whether this will scale or now. One test site called e-Kcollectors has five sites feeding ontologies and quickly reached 50 top levels. So we could see so many ontologies and categories that the diminishing utility is rapidly reached.