Some of the features of this wiki:
Restructured Text is fully supported, and is extended in a way that is natural both for a wiki and for reST. Any links (like link_) that cannot be resolved internally in a document are considered to be wiki links to other pages on the system.
Because Restructured Text is a full-featured markup language, it does not cut corners based on a limited initial domain. It is not wiki-specific, and does not produce output that is in any way wiki-specific or of inferior layout. It does not use WikiNames.
WYSIWYG editing using Xinha. Includes a custom button to insert intra-wiki links.
The wiki supports using an external editor to edit content, to avoid through-the-web editing (browser textareas are poor for editing).
The wiki also supports the Universal Edit Button to edit content.
Text and title searching. Incremental page searches when finding pages to link to.
"Pages" can be of any type, including binary and image files (e.g., PDF).
XML RSS feeds for new pages and recent changes available for updates of the site.
All edits are versioned, and a history is kept. You can compare versions.
Administrators can easily rollback changes, remove old versions or remove user accounts. This helps a lot in cleaning from spam.
You can plug in your own captcha for creating user accounts. Currently there are captchas suited for Python programmers and musicians. Robots and stupid spammers will not be able to create user accounts, but for serious users it is easy.
The wiki can be published to static files, with URLs that will be analogous to the in-wiki URLs. You can provide your own Cheetah template to control the page content. Pages are regenerated as their wiki equivalents are edited. Links to pages that do not exist are removed from the static published content; creating those pages will regenerate any pages that contained the previously dangling link.
These pages can be statically published both locally, and to an SFTP or FTP account.
Multiple domains can be served off a single installation; both aliases (e.g., http://wiki.webwareforpython.org and http://wiki.w4py.org), or entirely separate sites (http://wiki.cheetahtemplate.org).
Configuration is done through a simple .ini file. Configuration can be global or domain-specific. One possibility would be to configure two domains to point to the same content, but for one domain to be marked read only. As this wiki becomes more configurable, this could allow for it to grow into more CMS-like features.
All data is stored in simple text files; pages in .txt files, metadata in rfc822-style files, RSS in its own RSS file (as the canonical source of data), configuration in an .ini file. There is no binary data and no pickles (except for some indexing which is done in bdb files, which are not canonical sources of data -- you can delete and regenerate those files at any time). This makes upgrading easy -- you only need to update the code, not the data.
Doesn't require any special tools, like RCS. It is self-contained (beyond the Webware requirement, and modules in the standard Python library). Though it has grown some other optional libraries.
The logic for the wiki is separated from the logic for the interface. The wiki is written in independent modules which are not bound to Webware or any particular interface.
Another way of relating wiki pages to each other. More on its page. This is the beginning of blog functionality.
Comments are a certain kind of related page -- both individually addressable and editable, but are also shown inline in the commented-upon page.