Why we like MediaWiki

Mediawiki differs from most complex, database-driven Content Management systems: Rather than modeling the knowledge domain in the database, it is based on a flat, document based format supporting a flexible syntax for semantic and formatting markup. This format has been maintained stable for a decade and recent developments (SMW) even bring the format to the front of the semantic web, supporting RDF and OWL based expressions and automated reasoners. Current experience with Drupal shows the potential dangers inherent in module-specific tables. If a Drupal module is no longer supported in a new version, data remain inaccessible except for direct access to the mysql database.

In contrast, the Mediawiki model is layered rather than modularized. A very primitive plain text editor and a simple html rendering model always guarantee access and the ability to edit all data. Additional layers and extensions can add convenience and functionality to this, but the information layer itself cannot break as long as the basic software remains supported. Due to the huge investment into content present in the Wikipedias, it is unlikely that Mediawiki data will ever be left without a convenient upgrade path. This makes a strong case for Mediawiki as a long term manageable information management system.

Mediawiki is designed for a specific use case of working openly in large collaborations. The management model is very flat, and while write access control can be managed (anonymous write access is not planned for any of our deployments), read access is not meant to be managed. This makes it unsuitable for projects involving secure or sensitive information. At the same time it is socially very acceptable to be working among peers rather than voluntarily submit to a deep management and restriction scheme. It also make management scalable to very large collaborations.