The \textperthousand
command is available both with the T1 encoding (used by classicthesis
) and the TS1 encoding.
However, the Palatino font loaded via the mathpazo
package by classicthesis
hasn't the required glyph: in the T1 encoding \textperthousand
is built by adding a small zero next to %
and the small zero is missing (a black square is used to show this).
However the Palatino text companion font has the glyph \textperthousand
, so all you need to do is to load textcomp
: add
\usepackage{textcomp}
to your document preamble.
Note that \textperthousand
is not legal in math mode and produces a warning. You can avoid it by using
\mbox{\textperthousand}
or, better, by loading also amsmath
and using
\text{\textperthousand}
You may want to define a variant command that works both in text and math mode:
\usepackage{textcomp}
\usepackage{amsmath}
\DeclareRobustCommand{\perthousand}{%
\ifmmode
\text{\textperthousand}%
\else
\textperthousand
\fi}
Here, I just scaled the \square
in the vertical direction by a factor of 1.5 times the width, and called it \tallqed
. The \smash
prevents it from affecting line spacing. EDITED to reduce size. Note that first argument of \scalebox
is the horizontal scaling, while 2nd (optional) argument is the vertical scaling. These can be adjusted to suit.
\documentclass{article}
\usepackage{amssymb,graphicx}
\def\tallqed{\smash{\scalebox{.75}[1.125]{$\square$}}}
\begin{document}
\noindent In view of the choice we made for the orientation of $\partial D$, we conclude that
\[\int_{\partial \mathbf{D}} \iota^*\omega =
(-1)^n \int_{\mathbb{R}^{n-1}} a_n(\cdot,\cdot, \dots, \cdot, 0).\]
This completes the proof of the theorem. \tallqed
\end{document}

WChargin correctly points out that the symbol stretch makes the box border thickness non-uniform on the sides compared with the top/bottom. If that is an issue, the problem can be remedied with a slightly altered definition, by \ooalign
ing two of the stretched \square
s with a slight kern.
\documentclass{article}
\usepackage{amssymb,graphicx}
\def\tallqedX{\smash{\scalebox{.75}[1.125]{$\square$}}}
\def\tallqed{\ooalign{\tallqedX\cr\kern.2pt\tallqedX}}
\begin{document}
\noindent In view of the choice we made for the orientation of $\partial D$, we conclude that
\[\int_{\partial \mathbf{D}} \iota^*\omega =
(-1)^n \int_{\mathbb{R}^{n-1}} a_n(\cdot,\cdot, \dots, \cdot, 0).\]
This completes the proof of the theorem. \tallqed
\end{document}

Best Answer
A solution with TikZ: