Comment by immibis

We don't accept this from any other file format - why is HTML different? For example, if I include random blocks of data in a JPEG file, the picture is all broken or the parser gives up (which is often turned into a partial picture by some abstraction layer that ignores the error code) - in both cases the end user treats as completely broken. If I add random bytes into a Word or LibreOffice document I expect it not to load at all.