Thanks for the link! Happy to answer any questions.
A couple details that didn't make it to the blog post:
* I tried some smoother connection approaches than just intersecting the lines, but couldn't get it to look good in any way. Lots of very ugly iterations in between.
* At one point I/Codex needed to adjust some of the kerning and glyphs to avoid tiny line artifacts between connected characters, on visible at certain resolution/size combos. Not sure if that was the standard solution either.
This is pretty cool. one thing that jumped out of me though was that the colours did not alternate consistently.. for example you have W and X as BB/GG variants but why not several others? (eg A, B, D, H, K, M, N, O, P, Q, R, U, V & Y)
There are actually variants of BB and GG for all the letters as well in the font, but the calt rules will never select them in regular text. But a program could yoink them out for use.
If you reliably pick only BB or GG, the resulting text looks... stripey with only black or green between letters.
There's probably a more complicated way to cycle over color variants so that there's more variation in appearance too. Something to think about
A couple details that didn't make it to the blog post:
* I tried some smoother connection approaches than just intersecting the lines, but couldn't get it to look good in any way. Lots of very ugly iterations in between.
* At one point I/Codex needed to adjust some of the kerning and glyphs to avoid tiny line artifacts between connected characters, on visible at certain resolution/size combos. Not sure if that was the standard solution either.
If you reliably pick only BB or GG, the resulting text looks... stripey with only black or green between letters.
There's probably a more complicated way to cycle over color variants so that there's more variation in appearance too. Something to think about