{\LaTeX{} source from word processors} {Bart Childs} {Hennings' \CTAN\ survey is a good starting point when considering projects implied by the title of this article. I find it a fair view of most related packages. He suggests having one of two goals: converting the \textit{document structure} or converting the \textit{appearance}. My goal is neither of these: it is producing \LaTeX{} source that is clean and therefore maintainable. This is in keeping with Knuth's original goals in producing \TeX{}: graphic excellence and a document convenient for archiving. Structure and appearance are important but neither of these in the word processor format are as important as clean \LaTeX{} to me. My current system is a hybrid. I use the common word processor exchange format\Dash{Rich Text Format}, the OpenOffice Writer package and its \texttt{Writer2LaTeX} application, and macros for the Emacs editor written in Elisp. The test cases for this system are books on rotordynamics, a \acro{CS/1} programming text, a memoir on a friend's life including significant text fragments in the Czech language, and a novel that includes three love triangles. Even the worst case of significant mathematics formatting done in Word Perfect is tractable (I did not say easy). The most surprising problems are due to the limited skill of the users of their word processors. }