The Parsoid Confidence Framework is a mechanism for transparently evaluating the readiness ofParsoid Read Views on diverse wikis. This guide describes the goals of the framework and the metrics used in order to allow insight into the decisions of the Content Transform Team to deploy Parsoid as the default wikitext parser on a given wiki.
An overview andsummaries of the reports are atParsoid/Parser Unification/Confidence Framework/Reports.
Detailed results for current and planned deployments are within sub-pages:
Parsoid began as part of theVisualEditor project, and has been used in production as part of VisualEditor, mobile apps, and tools likeFlow andContent Translation since 2012. TheConfidence Framework was developed alongside the first deployment ofParsoid Read Views in 2024, part of theParser Unification project. As described onour updates page, our first deployments were to Wikivoyage wikis, followed by Incubator wiki and newly-created wikis. As the roll out of Parsoid Read Views progresses, the Confidence Framework will continue to be developed to reflect newly-uncovered challenges.
In particular, we plan to incorporate performance metrics into the Confidence Framework in the near future to ensure it properly reflects Parsoid readiness for large-page-view wikis.
Although the framework is built from numeric metrics, we do not blindly set cutoffs. Instead, the metrics are used to prompt product questions and investigation when they seem atypical.
These are the current metrics we use for decision making:
Based on the combination of the metrics we have an answer for enabling Parsoid Read Views for every wiki:
With that we will be able to look at the reasons for not being confident with Parsoid and focus on its characteristics and architecture with a Product Strategy lens to mitigate the gaps that affect our confidence to deploy Parsoid for Read Views.