Page MenuHomePhabricator

[SPIKE] Investigate technical feasibility of surfacing "Add an image" suggestions within Suggestion Mode
Closed, ResolvedPublic

Description

This task involves the work of investigating what technical work would be involved in surfacing Add an image suggestions within the VisualEditor's Suggestion Mode.

Learning objectives

  • 1. What technical work would need to be done to enable the experience described in Requirements below?
  • 2. What (if any() technical work would the Editing Team need to depend on other teams to complete?
  • 3. What level of effort/complexity would you assign to the technical work the Editing Team would need to take on to enable the experience described in Requirements below?

Requirements

For the purposes of this investigation, surfacing "Add an image" suggestions within Suggestion Mode would result in the following user experience:

  1. When people open the visual editor on a wiki where they've enabled Suggestion Mode and an "Add an image" suggestion is available
  2. Present that image suggestion in the way that all other suggestions are surfaced in Suggestion Mode
    • E.g. presented in the side-rail (desktop) and gutter (mobile), using a card design that is consistent with other suggestions, with the content the suggestion is relevant to highlighted, etc.
NOTE: "Add an image" suggestions surfaced in Suggestion Mode would need to use the same configuration settings that volunteers have already set at that wiki for "Add an image"

Done

  • Answers to all "Learning objectives" are documented on this ticket

Event Timeline

Our technical dependency would be an API endpoint from Growth that could tell us which images in which sections to suggest for a given article.

From there, showing a check card wouldn't be too hard, but most of the complexity is in the add-image workflow, as the image will be inserted without a caption. We have a few options:

  1. Place the cursor inside the empty caption and hope the user fills it out
  2. Show an "empty image caption" edit check after inserting the image
  3. Come up with a completely custom guided workflow like Growth uses, although I think this goes against our principle of building checks in a scalable and easy to maintain manner.
ppelberg claimed this task.

Our technical dependency would be an API endpoint from Growth that could tell us which images in which sections to suggest for a given article.

From there, showing a check card wouldn't be too hard, but most of the complexity is in the add-image workflow, as the image will be inserted without a caption. We have a few options:

  1. Place the cursor inside the empty caption and hope the user fills it out
  2. Show an "empty image caption" edit check after inserting the image
  3. Come up with a completely custom guided workflow like Growth uses, although I think this goes against our principle of building checks in a scalable and easy to maintain manner.

Wonderful. Thank you, Ed. I've translated these learnings into the newly-created T416043.