suggest that Wikipedia would be better if it's physics articles had interactive simulations or that stack overflow's failed docs initiative should make room for working examples and get downvoted into oblivion
I think you missed a vital reason and that is portability and sharability. No one needs to install anything special to read or edit a text file.
It's the digital equivalent of paper and pen when you think about it and that makes it a popular platform for everyone to be able to consume it.
Plus, back when we were all old and dinosaurs roamed the Earth, the text file was the lightest/smallest way to transfer data (read: ideas) between computers.
The practice/behaviour saturated the industry so much that many readmes of today are still on text files.
Maybe there's a correlation to the real world, where we're definitely more apt to consume mediums that don't have superfluous data points than the words/ideas that they're meant to convey? Maybe I'm just talking out of my ass?
Either way, plain-text resources are the easiest to create, share, and consume because there's almost a universality in the standard for text files (except the EOL and CRLF but that's more of an inconvenience than anything).
Techies like plaintext because it's free of superfluous bullshit. It's efficient. It's flexible.
As for interactive simulations, I'm all for them. I love the work of Bret Victor, or Nicky Case, or other stuff in the category of "explorable explanations". Those people do it well. But if it's not executed well, it's worse than not being done at all.