Robert Berkman has a post on his blog about Serph and TagFetch — two new metasearch engines that will search multiple blog search engines and tag aggregators (including Technorati, Flickr, Digg, YouTube, Google Blog Search, Bloglines and Newsvine).

The growth of these sites demonstrates the need for a metadata strategy for the web — which cannot possibly come from one central entity. Instead, it has to come from the users who create content, thus making distributed tags the only feasible metadata solution.

Tied to this is the growth of niche search engines and the ability to create your own search engine to search niche content (which Google has already delivered).

It would seem that as the amount of metadata attached to content on the web grows, there is a form of metadata that is evolving to categorize metadata itself (should we call it ‘meta-metadata?’).

Thinking about how to organize the increasing amount of metadata that is generated by content creation on the web will be a growing part of how we make sense of online information.

These growing tag clouds, however, will soon need someone to start identifying interrelationships between tags. It’s only a matter of time (perhaps it’s already been done) until someone creates an ontology to relate tags to each other.

Imagine a web where you could see not only visualize the popularity of terms, but also determine the relationships between those terms.

The image included in this post is what showed up on Flickr (found through TagFetch) in a search for ‘metadata’ — perhaps we’ve already started to visualize what tags can represent.