[go: nahoru, domu]

Page MenuHomePhabricator

Generic framework for implementing suggestions in VisualEditor
Open, Needs TriagePublic

Description

We (Growth-Team) are currently working on a project to guide a user in approving/disapproving new links in an article based on recommendations from a machine-learning library. One of the design concepts (concept A in T253020) explores some ways we might do this.

As we think about building, though, @Tgr pointed out that there would be a lot of value in establishing a generic framework (plugin?) for VisualEditor to allow interaction with suggestions in a uniform way.

The recommendation UX in VisualEditor. It's a fairly generic concept that works for most machine-generated fixes, we'll have more of those (spelling fixes, for example, or adding citations), it makes sense outside the recommendation context as well (e.g. "spell-check my edit" would be a useful generic functionality in the editor; most mainstream document editing software has it, with a similar look and feel); it would make sense to integrate it into the standard UX library.

While we plan to start with add link, later this fiscal year we plan to add a structured task to guide a user in adding suggested images. So we should probably start thinking about how to do both of these in a consistent way that integrates nicely with VisualEditor.

I'm creating this task so we can start discussing ideas.

Event Timeline

Dropping by to link to T95500 which I think is relevant and has come up in the conversation we've been having on the Editing Team about how the visual editor might be enhanced to help people learn the social conventions (read: policies and guidelines) and judgement necessary to become productive editors (see: T265163).

Removing from the list of tasks needed for our initial release, though we should revisit this again in the context of the #image-recommendations work