# Why does mathdesign result in a huge 'cases' environment brace in Tufte documents?

I'm getting strange results when I use the cases environment in Tufte-LaTeX documents with the mathdesign package. Instead of

I get

Is there something I can do to get rid of that huge brace?

\documentclass[]{tufte-handout}

\usepackage{mathtools}

%The following appears to be the source of my problems; removing this line produces an acceptable result
\RequirePackage[bitstream-charter,expert,euro]{mathdesign}

\begin{document}

$a(1,j)=\begin{cases} r^{j-1} & ;j\le n\\ f\cdot r^{j-(n+1)} & ;j>n \end{cases}$

\end{document}

-
The fix is simply to set \delimiterfactor=912. That seems to be the largest value that solves the problem. Tufte-LaTeX must set it to something larger. – raxacoricofallapatorius Jul 8 '12 at 22:21

The big brace you get is built from smaller pieces and that's the minimum size as the following minimal document shows:

\documentclass{article}
\usepackage[bitstream-charter]{mathdesign}

\begin{document}

\def\test#1{$\left\{\rule{0pt}{#1pt}(#1\textrm{pt})\right.$ }

\noindent\test{6}
\test{8}
\test{10}
\test{12}
\test{14}
\test{16}
\test{18}
\test{20}
\test{22}
\test{24}
\test{26}
\test{28}
\test{30}

\end{document}


Up to the 18pt case the brace is a unique glyph. Beyond that point TeX must build the brace from smaller pieces and in the case of the Charter math fonts provided by mathdesign this has an unfortunate gap.

There's a fundamental difference in behaviour between tufte-book and article: the former sets the normal baselineskip to 14pt instead of the 12pt of the article class. So, with tufte-book, the array that contains the two cases is

\vbox(28.55986+5.04002)x83.62628


while with article it is

\vbox(24.47987+4.32002)x80.47983


So it's a bit wider both horizontally and vertically. Indeed, if you try

$a(1,j)=\begin{cases} r^{j-1} & ;j\le n\\[4.8pt] f\cdot r^{j-(n+1)} & ;j>n \end{cases}$


with the article class, you'll see the same behavior as with tufte-book.

-
How do I avoid tipping over that 18-20 pt gap in the example above? – raxacoricofallapatorius Jul 3 '12 at 12:06
@raxacoricofallapatorius In that particular case, inputting r^{\smash{j-1}} avoids jumping to the next size. – egreg Jul 3 '12 at 12:44
Is there a general way to override LaTeX's judgment on the side of brace that should be used? And (perhaps this is a meta equation) is there a way to confirm that this is a logged bug, or get some sense of when it's likely to be fixed? – raxacoricofallapatorius Jul 3 '12 at 13:06
@raxacoricofallapatorius You should look at \delimitershortfall and \delimiterfactor in the TeXbook or TeX by Topic. But I'm afraid that this is a limitation of the specific font, not of TeX. – egreg Jul 3 '12 at 13:16
@raxacoricofallapatorius It's an option: \\[<dimen>] works also in cases. – egreg Jul 3 '12 at 14:06