**Type of activity:** Unconference session
**Main topic:** Handling wiki content beyond plaintext
== The problem ==
ProofreadPage, a MediaWiki extension that is intensively used on Wikisource, manages two specific content models:
* Page: pages that powers the proofreading interface that are an extension of regular Wikitext pages and composed of 3 connected wikitext blobs. e.g. https://en.wikisource.org/wiki/Page:Tragedies_of_Euripides_(Way_1894)_v1.djvu/234
* Index: pages that stores metadata about a book and its proofreading (author, table of contents, mapping between scan file pages and book pages...). e.g. https://en.wikisource.org/wiki/Index:Tragedies_of_Euripides_(Way_1894)_v1.djvu
These two models are currently stored as wikitext matching a specific format (for Page: pages the first and third wikitext blobs are serialized inside of <noinclude> and for Index: pages, the content is inside of a template serialization), mostly because of historical reasons.
What could be done in order to improve storage of these contents?
== Expected outcome ==
Have a plan of migration of these content models to something more healthy (Wikibase for a part of Index: pages content?).
== Current status of the discussion ==
* ContentHandler interface have already been implemented for Page: pages content
* modification have been done to Parsoid in order to allow proper handling of this kind of contents (but Page: pages are not yet cleanly supported by Parsoid): https://gerrit.wikimedia.org/r/#/c/295707/
== Links ==
* https://www.mediawiki.org/wiki/Extension:Proofread_Page