How to provide preferences for hyphenating mixed-language words

Question

Let's say we want to typeset long two-language compounds. We prefer to break such words right after the first-language word is over and before the second-language words starts; if the results get ugly (say, overfull boxes), we wish to allow for breaking also within the single-language parts.

Consider, e.g., Pipes-and-Filters-Architektur, Map/Reduce-Konzept, Assume/Guarantee-Spezifikation, and Quicksortprinzip — all of the kind English+German.

Here are our attempts so far:

\documentclass{article}
\usepackage[USenglish,ngerman]{babel}
\babelprovide[hyphenrules=ngerman-x-latest]{ngerman}%%% according to the documentation of dehyph-exptl.
%\showoutput
\begin{document}
\showhyphens{
\foreignlanguage{USenglish}{Pipes\penalty10000-\penalty2\hskip0pt and\penalty10000-\penalty2\hskip0pt Filters}\penalty10000-\penalty1\hskip0pt Architektur % Pipes-and-Filters-Architektur

\foreignlanguage{USenglish}{Map\penalty10000/\penalty2\hskip0pt Reduce}\penalty10000-\penalty1\hskip0pt Konzept % Map/Reduce-Konzept

\foreignlanguage{USenglish}{Assume\penalty10000/\penalty2\hskip0pt Guarantee}\penalty10000-\penalty1\hskip0pt Spezifikation % Assume/Guarantee-Spezifikation

\foreignlanguage{USenglish}{Quicksort}"-prinzip % Quicksortprinzip
}
\end{document}

1. Would this rock? We get the output (hyphens with \showhyphens and penalties with \showoutput) which seems right, but I am still not sure.

2. The above typesetting seems quite cumbersome; is there an easier way to reach the same goal? Is \penalty10000 really necessary between a word and a slash or a dash?

3. How to say in \foreignlanguage{USenglish}{Quicksort}"-prinzip that we prefer breaking between the English part and the German part rather than breaking inside any of the two parts?

Answer 1

You've told us that your document is almost entirely in German, apart from "a few English (compound or simple) words, a few English abbreviations, a few short English phrases, a few mixed English-German compounds, and some full English sentences". I take it that you know how how to deal effectively (within the babel framework) with short English-language phrases. Abbreviations and acronyms should not (ever?) be allowed to be hyphenated, right? The main issue, then, is how to deal with the mixed English-German compound words, of which you've provided four choice examples: Quicksortprinzip, Pipes-and-Filters-Architektur, Map/Reduce-Konzept, and Assume/Guarantee-Spezifikation.

My view is to treat these cases as if they were (mostly) German-language constructs. Thus, do use \slash instead of / to allow line-breaking, and do make use of the babel-german "= shortcut to allow hyphenation of the components of compound hyphenated words. Finally, use \babelhyphenation[ngerman]{....} to indicate where not to hyphenate certain words. E.g., left to its own devices, babel-german "finds" tso hyphenation points in both assume and reduce; I suggest suppressing hyphenation entirely for these two words.

I've used LuaLaTeX to compile this test document, mainly so that the hyphenation of the first word in a paragraph is enabled .

% !TEX TS-program = lualatex
\documentclass{article}
\usepackage[english,ngerman]{babel}
%% suppress hyphenation for selected English words in a German-language context
\babelhyphenation[ngerman]{guarantee filters reduce assume}

\setlength\parindent{0pt}
\setlength\textwidth{1sp}

\let\oldslash\slash
\renewcommand\slash{\oldslash\hspace{0pt}}
\begin{document}

Quicksortprinzip

\smallskip
Pipes"=and"=Filters"=Architektur

\smallskip
Map\slash Reduce"=Konzept

\smallskip
Assume\slash Guarantee"=Spezifikation
\end{document}