Background
We've decided we will scale the image browsing feature, starting with just (1) the image carousel and (2) detail view ingressing from the carousel
- based on data from our test https://people.wikimedia.org/~jiawang/reader_growth/image_browsing/image_browsing_usage_analysis_report.html
- based on team discussions: (1) https://docs.google.com/document/d/102ZNCBtP8IOOtcDpqA9G8k29jzugrk07fpEP8ha3uoc/edit?tab=t.0 (2) https://docs.google.com/document/d/1FAsXKhrrUU5z2AtWwMbgbag5_NBWOdO-Jme0YL-9R3E/edit?tab=t.0
The next step, covered in this task, for us is a technical exploration that documents options (e.g., leveraging mediaviewer, something new, other wonderful surprise option???) of how to scale the image carousel and detail view components to all logged-out readers and considerations for deciding between them.
Inputs & Assumptions
- Logged-out users only
- Mobile web only, but please call out tradeoffs in immediate effort, long-term maintenance, etc. that might cause us to consider including desktop web in this phase too
- Not scaling Visual Table of Contents (VTOC) and Images from Other Wikis for now
- Making the following design improvements (if you need mocks to clarify any of these, let @JScherer-WMF know)
- - Default to seeing all of the image
- - Robust pinch to zoom and swipe to pan affordances
- - Jettison the vtoc and the images from other wikis
- - Make it clear that the horizontal gallery is scrollable
- - Update visual design on captions
- - Update image types/tags we want to exclude, e.g., signatures and data viz
- - Better autocrop
Open Questions to Be Covered
- What are all the things mmv can do on mobile? Corollary: what are all the things mmv can do on desktop?
- If we make changes to mmv for this feature, will they be disruptive enough that we need to add an account-based preference setting?
- How much work is it to add an account-based preference setting for image browsing?
- [Nice to have] What work would be required to build in such a way that the top-of-page real estate becomes a feature-agnostic container for future tests?