People slowly learned to use their keyboards for proper typography?

If the graph is not just using made up data, then it also measures absolute em-dashes, not em-dash density, so possibly could be an increase in interest in contributing to the kernel?

I assume it's due to phone keyboards.

The OP graph is just absolute count on the y axis, so I also assumed it was basically measuring LoC, but looking at this chart, the em-dashes do appear to be definitely more of a modern phenomenon compared to raw LoC:

Maybe some policy/tooling change around 2014? A guess is perhaps something to do with allowing UTF-8 in source files, but that's a pretty long shot.

I wonder how many em dashes are real vs interpreted.

I love an em-dash, but I have never typed one on purpose; they are all just two minus in a row and I use them wrong -- with a space around (*inspect source to see that lemmy translates for me). I abuse them like I abuse semicolons. They are a typographical tick that indicates a pause or pivot when I translate my thoughts to text.

I can imagine a human manuscript editor turning punctuation like a comma into an em-dash when it makes sense, but I can't imagine a developer ever specifically typing out one -- with the need to either know or look up the unicode for it. The only relevant place would be a comment, but even then IDE's are not generally known for their tendency to replace typed characters with something else.

Maybe some tool in the pretty-print/code-formatter ecosystem added that as a feature? But var-- is so common that I can't imagine that'd be a good idea, even in a comment.