Multilanguage Documents with TeX

Today I have been working on a LaTeX document which was supposed to produce a document in two versions, one in Italian, another in English. It was a technical report, so a lot of stuff, but not so much text.

The classic way to approach this is to have the document written in one language, and then translate it into the second. This can be fair, until you have to apply changes. Editing two documents in a row can be too much effort-consuming.

Luckily enough, $\LaTeX$ is a programming language, before an office automation tool. So we can use conditional variables to fulfill this task. So, the skeleton of my source looks like this:

\documentclass[10pt]{article}

\newif\ifit
\newif\ifen

\newcommand{\langit}[1]{\ifit#1\fi}
\newcommand{\langen}[1]{\ifen#1\fi}

\entrue
%\ittrue


\ifen\usepackage[english]{babel}\fi
\ifit\usepackage[italian]{babel}\fi

\begin{document}

\langen{This is English text}
\langit{Questo è del testo in italiano}

\end{document}

Basically, I declare two conditional variables, \ifit and \ifen which select which language code must be generated. By uncommenting either \entrue or \ittrue, the source will produce the output in English or in Italian.

To ease the task of writing the TeX source, two new commands are declared: \langit and \langen which accept one parameter (i.e., the text in the corresponding language) and output it only if the corresponding conditional variable is set.

Additionally, depending on which conditional variable is set, the babel package is loaded with the corresponding language.

This allows to work on one single TeX source (which decreases the maintainability effort), but allows to produce documents in multiple languages. Adding new languages is just a matter of creating new variables and commands.

Another interesting this, which does not appear in this example, is the charset. Some languages have different charsets, so it would be interesting to set it accordingly. For the italian case, it would entail adding before the document's begin, the following code:

\ifit\usepackage[utf8]{inputenx}\fi