That seems intuitive, though. EM radiation is either reflected or absorbed, and optimizing for that requires both a pretty complex understanding of RF behavior and generally knowing that materials are generally radiopaque and radiolucent at different frequencies and wattages.
Sometimes we're trying to keep things (eg- information) outside from getting in, and other times we want to prevent things inside from getting out. There are practices to optimize for both that don't rely on "blocking".
By interfaces yes, but it can also be cancelled out through destructive interference as a side effect of reflection, which is my theory of how a "big ball of crumply aluminium" is so effective compared to less chaotic solutions.
Rough surfaces increase reflection in non-specular directions and decrease it in the specular direction. I have never heard that it would facilitate destructive interference.
Sometimes we're trying to keep things (eg- information) outside from getting in, and other times we want to prevent things inside from getting out. There are practices to optimize for both that don't rely on "blocking".