Title: How does Nukumi2 work?
Page: how-it-works

Nukumi2 can work in two major ways: dynamically (using Webrick) and
statically (rendering to files).  However, these two ways only differ
marginally.  We'll assume you to use dynamic usage for now.

On startup, Nukumi2 will parse and load the `config.rb` file of the
site.  This file essentially contains all configuration options used.
For example, it defines title and webmaster of the site.

It also, and that is more important, defines the backends used as well
as the skeleton TopicTree.

Next, Nukumi2 will query all backends on available entries and collect
them.  Entries now insert themselves to a TopicTree.

Nukumi2 is now ready to publish pages, and enters Webrick.

Upon a request, Nukumi2 splits the URL into two parts: the topic and
the flavor.  While the topic specifies the content, the flavor
specifies the form.

The topic is now matched against the skeleton topictree to determine
the view to use.  Often used views are `Last` (which returns the last
*n* Entries in reverse chronological order, this is what most people
use for the front page of the blog.), `All` (all entries about the
topic, for example for archives) and `Single` (exactly one entry, with
full content, e.g. for permalinks).  This view now gets called passing
the topictree and the topic and will usually (although it doesn't need
to) arrange and filter them in some order.

In the next step, the view gets inserted into a page (of a particular
flavor) that assembles all the parts of the final page using a
templating engine together.  Nukumi2 will now send this page back to
the browser/client.