2026-07-25 03:30 UTC
We need to output the determined glyphs onscreen at their positions. As we split text into lines we may need to simultaneously split it into differently-styled text. And we may need to calculate where to position these paragraphs based on how wide they might be & how tall they are for a given width, or vice-versa if we're dealing with vertical text.
Or perhaps we want to render to audio rather than visuals? That's more straightforward, but similar.
So how'd my hypotheticals handle it?
3/5?
Replies (1)
-
@alcinnz@floss.social 2026-07-25 03:40
When designing a hardware-browser I found that the string-centric CPU designed could easily choose fonts, resolve bidirectional text, & otherwise preprocess text for rendering. But when I designed a coprocessor to (concurrently) mathematically negotiate layout, I found it'd need to layout the text as part of that negotiation. Which in turn necessitated an extension letting it consult (small-ish) lookuptables, in addition to dangling the font's Binary Search Trees off the layout tree. 4/6?