The Open Directory Project creates a hierarchical, subject oriented WWW index where maintenance of categories is delegated to authors. It stands in a long tradition (measured in web lifetime).
The first implementation of this idea I ever saw was the WWW Virtual Library. This was in 1992, before Mosaic existed, before America discovered the Web.
Several categories existed with their own maintainers, but no software was available to support this.
The second such system was Oliver McBryan's 'Mother of all BBS': it allowed arbitrary users to add nodes and subtrees to a hierarchical namespace. It was meant to be completely self-organising. More of a software concept demo than anything else - with some of the Everything 1 feel to it. (At the same time, McBryan operated the WWWWorm, one of the first full text search indexes for the Web.) This was in 1994.
The approach only became popular when two Stanford students used the same ideas and created Yahoo!.
I like the idea of an Open Directory Index but I cringe at the idea of the duplication of effort involved in all these different hierarchies. When it goes commercial, it will just be the next Yahoo! and everything will start all over again.