People have distinct personalities. So do text fonts. I wonder whether we could combine together these two concepts.
When we look at certain text fonts, we usually agree that they look happy, or elegant, or frivolous, or foreboding,or any number of qualities that we usually associate with people. We could probably run a supervised machine learning algorithm that would train a font generator from labeled examples.
We could then, for any well described human personality, develop the perfectly descriptive font.
This might be a very useful thing to have. For example, in online text chats, we could immediately know what kind of person we are dealing with.
We could apply the same principle to written stories with narrative voices. As each person talks, the text font of their speech can be made to reflect deep aspects of their personality.
This was actually done a long time ago in the world of comic books. My uncle Abe Kanegson was a letterer for Will Eisner’s The Spirit series of comics back in the day. He would invent new fonts to match particular scenes, situations and personalities.
And now, all these years later, we can finally do that with computers!