Look and feel

• Links with shortcuts, arrangement depends on skin
• Skins: Different ways to present the site.
• User styles: Users can adapt the look and feel of the site through custom CSS on their user pages.
• "Stub" threshold: Users can see links to articles below a certain size rendered in a different color
• Printable versions of articles can be generated
• Auto-number headings in an article (optional)
• Intra-page Anchors
• Generate a table of contents for long articles (optional)
• Automatically turn ISBN numbers into links to an editable list of booksellers
• XHTML-compatible output (or darn close to it), tidy integration

Multimedia and extensions

• File upload feature allows to upload graphics or sound files
• Mathematical formulas using LaTeX syntax
• Automatic resizing of images using ImageMagick or libgd, simple syntax for image captions and image alignment
• WikiHiero for Egyptian hieroglyphs (needs separate installed components)
• EasyTimeline for time charts (needs separate installed components)

Keeping track of edits

• Watchlist Every page has a link "Watch this article for me".
• User contributions in the sidebar of each user page list all articles the user has worked on, according to the database.
• Extended recent changes with dynamic collapsing of edits to the same article and quick links to diff the edit, show the article history, show the user page, show the user talk page, or block the user (for sysops)
• "Related changes": View a filtered version of Recent Changes to the pages linked from the current page.
• Side-by-side diffs - the diffs are shown side-by-side, and changed portions of lines are highlighted, making it much easier to see what's what. Additionally, a diff is shown during an edit conflict so you can see exactly what you need to reintegrate.
• Real names. Users can (optionally) specify a "real name" they want to use for author credits.
• On-page credits. Administrators can enable an on-page paragraph giving credit to editors who've worked on a page.

Backend

• Database-driven (MySQL and PostgreSQL are both supported)
• Parser caching: rendered pages are cached to reduce redundant processing
• Output caching: complete page output can be optionally saved as static HTML files for serving to other clients; support for Squid proxies
• Cookie-based account and login system, but anonymous users can also edit pages
• All revisions of an article are stored (optionally compressed)
• RSS syndication available for many special pages
• RDF metadata