And, if it's a real website, with branching connections, the author(s) cannot anticipate how any visitor will move through its pages.

In practice, any visitor may encounter any website via the discovery of any page. From there, in-text links quickly develop a complex non-linear set of possible directions. The implication is that no index can really apply to the way that pages work together.

One can make thematic directories; but there can be no directory for the whole project — although the idea can be part of the project.

The organizational theory is sort of based on the self-ordering principle of sparsely-connected networks.

Does it work? Sometimes, in some ways; not always.

It's an experiment, ongoing.