How to make sections into sentence case

Question

I would like my sections (and potentially subsections) to be in "sentencecase" no matter how I type them. I am happy with the way the \MakeSentenceCase macro from the biblatex package converts text with arbitrary capitalization into sentencecase. I am having difficulty integrating this into the \section macro.

My first attempt was to redefine \section to be

\renewcommand{\section}{\@startsection{section}{1}{\z@}{-3.5ex \@plus -1ex \@minus -.2ex}{2.3ex \@plus.2ex}{\normalfont\Large\bfseries\MakeSentenceCase}}

but this fails because the first token (or something like that) to \MakeSentenceCase is not the first token of the text of my section heading, but rather related to the numbering and indenting. The following method works

\documentclass{article}

\usepackage{etoolbox}
\usepackage[backend=biber]{biblatex}

\makeatletter
\let\sectionOrig\section
\renewcommand{\section}{\@ifstar{\@sectionStar}{\@sectionNoStar}}%
\newcommand{\@sectionNoStar}[2][\relax]{\ifdefstring{\relax}{#1}{%
  \sectionOrig{\MakeSentenceCase{#2}}%
}{%
  \sectionOrig[#1]{\MakeSentenceCase{#2}}%
}}
\newcommand{\@sectionStar}[1]{\sectionOrig*{\MakeSentenceCase{#1}}}%
\makeatother

\begin{document}
\section{sections SHOULD BE IN {SENTENCECASE}}
\end{document}

by completely redefining the \section command. I have tried to take care of the starred and unstarred forms and optional arguments, but I am worried that this approach is going to cause problems someplace else.

I guess I have 3 questions. First, is it reasonable to expect LaTeX to automatically handle the capitalization of sections. Second, is my approach reasonable. Third, is there a better approach (or an already existing package).

Answer 1

This is easy using the titlesec package. With the explicit option (p. 8, sec. 3.7), we can use an (I think) arbitrary transformation on the sectioning title stored as #1. With this knowledge, the rest becomes a piece of cake:

\documentclass[twocolumn]{article} % two-column used to get everything onto one page
\usepackage{biblatex}

\usepackage[explicit]{titlesec} % note the `explicit` option

\titleformat{\section} % The normal, unstarred version
    {\Large\bfseries}{}{2ex}
    {\thesection. \MakeSentenceCase{#1}}

\titleformat{name=\section,numberless} % The starred version; note the `numberless` key
    {\Large\bfseries}{}{2ex}
    {\MakeSentenceCase{#1}}

\usepackage{mwe}
\begin{document}
\section{I am a Section}
\lipsum[1]
\section{I am An Other section}
\lipsum[2]
\section*{I am an UnNumBeReD section with {ACRO}-nyms}
\lipsum[3]
\section{I am a Section}
\lipsum[4]
\section{I am An Other section}
\lipsum[5]
\section{More {ACR}-o-{NYM}s!}
\end{document}

(Note that this solution can be easily applied to the other sectioning commands by just replacing \section and \thesection appropriately.)

---

As for your other questions:

• as long as you are not expecting LaTeX to achieve/follow fancy casing rules (such as title-casing) for arbitrary languages, then you are fine. TeX knows the relationships between all of the letters (uppercase/lowercase) and can swap between them where necessary. So-called 'dumb' control of casing (e.g. capitalize first letter, lowercase the rest) is completely reasonable.
• Your approach is reasonable if it works well, but it is usually better to build upon an existing package (e.g. titlesec) for more functionality. (For example, if you had written your own \MakeSentenceCase, that may have been a bit too much work—a lot of code means a lot of room for small errors.)