I've long wondered how well using a DOM serializer that was a reasonable inverse of a browser would work. I feel like it appeals to my math-y "generalize everyt

I've long wondered how well using a DOM serializer that was a reasonable inverse of a browser would work. I feel like it appeals to my math-y "generalize everything" side, but might not work out well in the "many pages have special needs" real-world. This is pretty interesting. I'm interested to see how maintainable "pretty" HTML ends up being.