Make it interactive

Written for version 0.1.5.

Stage 5 of How it works. It starts only after you’ve approved every static view, and only if you ask for it.

Ask

“Now make the approved figure interactive.” If you want something specific, such as one view per tab, say so.

What you get

The approved figure stays the figure. Regions of it open to show more detail:

• On a desktop, hovering briefly over a region opens its full explanation; you can move the pointer into it, and click to keep it open.
• On a phone, tapping the whole region opens the same content.
• With a keyboard, each region is a normal focusable control with a visible focus outline, and panels close with Escape where that applies.

Only the region you selected is highlighted. Panels shouldn’t flicker, cover what you tapped, or close by accident.

If the whole figure is too small to read at phone width, the assistant adds a way to enlarge or pan it, or agreed crops of each section. It doesn’t swap your illustrated figure for a text list unless you agree. There is always a plain-text account of the content for screen readers and for anyone who can’t hover.

What stays fixed

The approved composition, illustrations, colours, relationships and content don’t change. A new view goes back through content confirmation and SVG approval; small refinements to one view don’t reopen the others.

Testing

The assistant tests hover, pinning, closing, keyboard, touch, zoom, the edges of the screen, scrolling and links, as far as it can. It tells you which tests were on a real phone or browser and which were simulated.