Hello Solan

We do not plan to use MediaWiki's data model (rich text) or editing concept (text based markup editing) for Wikidata. We will use a structured data model with form-based ajax enhanced editing.

We however do plan to implement Wikidata based on MediaWiki in order to be able to seamlessly integrate with the existing infrastructure and culture of wikimedia projects like Wikipedia. MediaWiki provides us with several important facilities:

* pages for discussions, policies, etc * user management, including SUL integration with Wikipedia et.al. * Robust versioning * import/export of dumps (backups, etc) * Scalable storage and caching facilities * Facilities for change tracking and notification * Permission management * Spam protection and filtering * ...and much more

...and it does all this a) for free and b) in exactly the way people who are already involved are familiar and comfortable with.

Basically, we need a software platform that provides all the necessary services for the community, not just storing and editing structured data. MediaWiki provides all this, so we will reuse it. We will implement the specialized functionality using MediaWiki's extension interfaces.

Some details on how we plan to handle structured data:

Basically, the primary copy will be serialized as JSON and stored in the database as a blob, using the same facilities currently used for wikitext. This allows us to using the existing infrastructure and processes for storage, versioning, load-balancing, backup, etc.

This of course doesn't provide us with an efficient way to query the data. In order to support queries, the current version of the data will also be stored in a (or multiple) secondary database - most likely a NoSQL database like hbase. MediaWiki already uses secondary data tables (albeit in MySQL) for many things it needs quick access too, e.g. the categories assigned to wiki pages.

HTH

Daniel Kinzler

On 09.04.2012 11:55, Soslan Khubulov wrote:

Hello!

What about the engine of Wikidata? Do you think MediaWiki is good for structured data? I think MediaWiki should be greatly modified at least to be proper engine for Wikidata.

The best thing would be to create new engine specially for structured data. It would be also better for Wikitionary. Just remember what was Mediawiki created for. Storing marked up text pages. Mediawiki is good for encyclopedia but not for Wikitionary and Wikidata purposes.

Sincerely, Soslan Khubulov

---

Wikidata-l mailing list Wikidata-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata-l