not to mention obvious old-school approach xml->anything (including
ConTeXt). 100 percent reliable conversion of pdf to anything else is
impossible by definition, so better don't rely on any magic tools
which usually promise but really cannot deliver.