nested figures
This is an HTML accessibility guide on captioning a group of related images, using a castle-through-the-ages series as its running example. It recommends an outer figure with role="group" and a group figcaption at the top, containing a nested figure for each image with its own figcaption. This keeps every caption visible by default and lets you mark up details such as a French photographer's name with lang="fr". A second example shows the same structure with CSS to show or hide captions on hover or focus, and the guide warns against the anti-pattern of putting several images in one figure with captions in title attributes, which keyboard and touch users can't reach and which can't carry any semantic markup.
15 years go by, and he's still right...