Spike: Parsoid Usage Confidence Framework
Closed, ResolvedPublic
Actions

Assigned To

Authored By

	MSantos
	Oct 16 2023, 2:56 PM

Description

Background Information

See the current draft proposal in this wiki page.

The goal is to turn on parsoid on self-contained and safe places to develop the framework for ‘Parsoid Usage Readiness’.

What

The idea is to use the framework to actively learn with our roll-outs and identify the opportunities and gaps we need to cover without being framed by a latter of phases and specific milestones but instead, evolving the product strategy for the Parsoid Product Domain (draft proposal) for a long-term strategy that also takes into account immediate goals and committed work.

How

Research how the current metrics can fit into this framework

Acceptance Criteria

We have a set of metrics to look before rolling-out Parsoid Read Views in the Officewiki or Wikitech DiscussionTools

Related Objects
Search...

View Standalone Graph

This task is connected to more than 200 other tasks. Only direct parents and subtasks are shown here. Use View Standalone Graph to show more of the graph.

Status	Assigned	Task
		· · ·
Resolved	cscott	T355374 Use Parsoid for DiscussionTools on wikitech
Resolved	cscott	T355566 Use Parsoid for read views on OfficeWiki by default
Resolved	MSantos	T348998 Spike: Parsoid Usage Confidence Framework
		· · ·

Event Timeline

MSantos created this task.Oct 16 2023, 2:56 PM

Restricted Application added a project: Parsoid. · View Herald TranscriptOct 16 2023, 2:56 PM

Restricted Application added a subscriber: Aklapper. · View Herald Transcript

MSantos triaged this task as High priority.Oct 16 2023, 2:56 PM

Some initial thoughts:

We don't have any active linter categories that impact parsoid read views -- but if we introduced any going forward, then yes, this would be a factor.
We can do wiki-specific visual diff testing for a subset of pages sampled from the wiki, but we cannot do this for private wikis right now. So visual diff testing for officewiki is currently difficult.
For initial rollouts (wikitech DT, officewiki), or even low-scale opt-in, performance may not be a consideration, but default enabling of read views on anything but the small wikis might need to factor in performance metrics.
For opt-in rollout, extension coverage gaps might be acceptable (but # of pages affected may be a metric?), but default enabling of read views will probably need 100% coverage OR a mechanism to opt-out affected pages (once again, # of pages that need to be opted out may be a metric).