% \iffalse
% -------------------------------------------------------------------
%
% Copyright 2001--2006, Daniel H. Luecking
%
% Mathdots may be distributed and/or modified under the conditions of the
% LaTeX Project Public License, either version 1.3c of this license or (at
% your option) any later version. The latest version of this license is in
%    <http://www.latex-project.org/lppl.txt>
% and version 1.3c or later is part of all distributions of LaTeX version
% 2008/05/04 or later.
%
% Mathdots has maintenance status "author-maintained". The Current
% Maintainer is Daniel H. Luecking. The Base Interpreters associated with
% mathdots are TeX and LaTeX.
%
%<*sty|tex>
%<*driver>
\ProvidesFile{mathdots.dtx}%
%</driver>
%<sty>\ifx\documentclass\UndEfInEd\input mathdots \expandafter\endinput\fi
%<sty>\ProvidesPackage{mathdots}
%<*sty>
[2014/06/11 v0.9 Improve and add various macros for dots in math.]%
%</sty>
%<tex>\def\mathdotsfiledate{2014/06/11}%
%<tex>\def\mathdotsfileversion{0.9}%
%<*driver>
\documentclass{ltxdoc}
\usepackage{amsmath}
\usepackage{mathdots}
\addtolength{\textwidth}{.5878pt}

\def\mytt{\upshape\mdseries\ttfamily}
\renewcommand\marg[1]{{\mytt \{#1\}}}
\renewcommand\oarg[1]{{\mytt  [#1]}}
\renewcommand\parg[1]{{\mytt  (#1)}}
\renewcommand\arg[1]{{\mytt\##1}}
\renewcommand{\meta}[1]{{$\langle$\rmfamily\itshape#1\/$\rangle$}}
\DeclareRobustCommand\cs[1]{{\mytt\char`\\#1}}
\def\prog#1{{\mdseries\scshape #1}}
\def\opt#1{{\sffamily\upshape#1}}
\def\mtt#1{{\mytt#1}}
\let\env\mtt
\let\file\mtt
\def\MD{\prog{mathdots}}
\renewcommand\{{\mtt{\char`\{}}
\renewcommand\}{\mtt{\char`\}}}
\renewcommand\|{${}\mathrel{|}{}$}

\makeatletter
\newcommand\bsl{{\mytt\@backslashchar}}
% Stupid lists!
\def\@listi{\leftmargin\leftmargini
  \parsep \z@ \@plus\p@ \@minus\z@
  \topsep 4\p@ \@plus\p@ \@minus2\p@
  \itemsep\parsep}
\let\@listI\@listi \@listi
\renewcommand\labelitemi{\normalfont\bfseries \textendash}
\renewcommand\labelitemii{\textasteriskcentered}
\renewcommand\labelitemiii{\textperiodcentered}
\leftmargini\parindent
% Stupid index!
\def\usage#1{\textrm{#1}}
\def\index@prologue{\section*{Index}\markboth{Index}{Index}%
  Numbers refer to the page(s) where the corresponding entry is described.
}
\def\IndexParms{%
  \parindent \z@ \columnsep 15pt
  \parskip 0pt plus 1pt
  \rightskip 5pt plus2em \mathsurround \z@
  \parfillskip-5pt \small
  % less hanging:
  \def\@idxitem{\par\hangindent 20pt}%
  \def\subitem{\@idxitem\hspace*{15pt}}%
  \def\subsubitem{\@idxitem\hspace*{25pt}}%
  \def\indexspace{\par\vspace{10pt plus 2pt minus 3pt}}}
\makeatother

\title{The \MD{} package\thanks{This file has version number
        \fileversion, last revised \filedate.}}
\author{Dan Luecking}
\date{\filedate}
\DisableCrossrefs
\CodelineIndex
\AlsoImplementation
\begin{document}
  \DeleteShortVerb{\|}
  \DocInput{mathdots.dtx}
\end{document}
%</driver>
%\fi
% \CheckSum{364}
% \CharacterTable
%  {Upper-case    \A\B\C\D\E\F\G\H\I\J\K\L\M\N\O\P\Q\R\S\T\U\V\W\X\Y\Z
%   Lower-case    \a\b\c\d\e\f\g\h\i\j\k\l\m\n\o\p\q\r\s\t\u\v\w\x\y\z
%   Digits        \0\1\2\3\4\5\6\7\8\9
%   Exclamation   \!     Double quote  \"     Hash (number) \#
%   Dollar        \$     Percent       \%     Ampersand     \&
%   Acute accent  \'     Left paren    \(     Right paren   \)
%   Asterisk      \*     Plus          \+     Comma         \,
%   Minus         \-     Point         \.     Solidus       \/
%   Colon         \:     Semicolon     \;     Less than     \<
%   Equals        \=     Greater than  \>     Question mark \?
%   Commercial at \@     Left bracket  \[     Backslash     \\
%   Right bracket \]     Circumflex    \^     Underscore    \_
%   Grave accent  \`     Left brace    \{     Vertical bar  \|
%   Right brace   \}     Tilde         \~}
%
% \catcode`\_=12
% \GetFileInfo{mathdots.dtx}
% \maketitle
%
% \begin{abstract}
% This package provides vertical dots and diagonal dots in math, slanting
% in either direction. It should work equally well in \LaTeX{} or plain
% \TeX{}. It improves on the default definitions of plain \TeX{} and
% \LaTeX{}. Similar improvements are provided for the triple and
% quadruple dot accents of AMS\TeX{}/AMS\LaTeX{}.
% \end{abstract}
%
% \StopEventually{\PrintIndex}
%
% \section{Introduction}
%
% Running \LaTeX{} or \TeX{} on \file{mathdots.ins} generates
% \file{mathdots.sty}, a \LaTeX{} package, and \file{mathdots.tex} a plain
% \TeX{} input file. If one puts \verb"\input mathdots.sty" in a plain
% \TeX{} file, it will cause \file{mathdots.tex} to be input. This is for
% backward compatibility.
%
% Unlike the commands \cs{ddots} and \cs{vdots} provided by \LaTeX{},
% the versions defined by \MD{} change size with \LaTeX{} size
% changing commands, as well as in \{sub[sub]\|super[super]\}scripts.
% After \MD{} is loaded, \verb"{\large$\ddots$}" will be larger and
% \verb"{\small$\ddots$}" will be smaller than \verb"$\ddots$" at normal
% size. Moreover, \verb"$2^{2^\ddots}$" will also produce a smaller
% symbol. This last example is very unlikely to be needed, but the version
% with the opposite slant is very common (and is provided by \MD{}).
%
% The \prog{amsmath} accents \cs{dddot} and \cs{ddddot} have similar
% problems that \MD{} attempts to correct when \prog{amsmath} is loaded.
%
% See Table~\ref{examples} for examples of the \MD{} commands at
% different sizes and in exponents. See Table~\ref{compare} for
% comparison, showing the results without \MD{}.
%
% \def\dott#1{$#1$}
% \def\dotts#1{$2^{#1}\quad 2^{2^{#1}}$}
% \begin{table}[htbp]
% \centering \renewcommand\arraystretch{1.4}
% \begin{tabular}{c|cccc}
% \multicolumn{1}{l}{\large With \MD{}:}\\[4pt]
% \multicolumn{1}{c}{\textbf{Command}}%
%                  &\textbf{Large}          &\textbf{normal}   &\textbf{scriptsize}
% &\textbf{in exponents}\\
% \hline
% \verb$\ddots$    & \Large\dott{\ddots}    & \dott{\ddots}    & \scriptsize\dott{\ddots}
% & \dotts{\ddots}\\
% \verb$\vdots$    & \Large\dott{\vdots}    & \dott{\vdots}    & \scriptsize\dott{\vdots}
% & \dotts{\vdots}\\
% \verb$\iddots$   & \Large\dott{\iddots}   & \dott{\iddots}   & \scriptsize\dott{\iddots}
% & \dotts{\iddots}\\
% \verb$\dddot{X}$ & \Large\dott{\dddot{X}} & \dott{\dddot{X}} & \scriptsize\dott{\dddot{X}}
% & \dotts{\dddot{X}}\\
% \verb$\ddddot{X}$& \Large\dott{\ddddot{X}}& \dott{\ddddot{X}}& \scriptsize\dott{\ddddot{X}}
% & \dotts{\ddddot{X}}
% \end{tabular}
% \caption{Dots at different sizes and in exponents.}
% \label{examples}
% \end{table}
%
% \begin{table}[htbp]
% \centering \renewcommand\arraystretch{1.4}
% \begin{tabular}{c|cccc}
% \multicolumn{1}{l}{\large Without \MD{}:}\\[4pt]
% \multicolumn{1}{c}{\textbf{Command}}%
%                  &\textbf{Large}                      &\textbf{normal}              &\textbf{scriptsize}
% &\textbf{in exponents}\\
% \hline
% \verb$\ddots$    & \Large\dott{\originalddots}     & \dott{\originalddots}    & \scriptsize\dott{\originalddots}
% & \dotts{\originalddots}\\
% \verb$\vdots$    & \Large\dott{\originalvdots}     & \dott{\originalvdots}    & \scriptsize\dott{\originalvdots}
% & \dotts{\originalvdots}\\
% \verb$\dddot{X}$ & \Large\dott{\originaldddot{X}}  & \dott{\originaldddot{X}} & \scriptsize\dott{\originaldddot{X}}
% & \dotts{\originaldddot{X}}\\
% \verb$\ddddot{X}$& \Large\dott{\originalddddot{X}} & \dott{\originalddddot{X}}& \scriptsize\dott{\originalddddot{X}}
% & \dotts{\originalddddot{X}}
% \end{tabular}
% \caption{Dots without \MD{}.}
% \label{compare}
% \end{table}
%
% \DescribeMacro{\ddots}
% The default \LaTeX{} (and plain \TeX) definition changes the size of the
% dots, and the horizontal dimensions, but not the vertical dimensions,
% giving them different slants at different sizes. Also, using the default
% version in 'scripts produces symbols identical to those produced
% elsewhere. Finally, the default takes the dots from the current text
% font, whereas the \MD{} package takes them from the current math symbol
% font, so they should match other dots in math mode. This is only really
% a concern in \mtt{17pt} sizes or larger where \file{cmr17} is scaled to
% get text dots, but \file{cmsy12} is scaled to get math dots.
%
% \DescribeMacro{\iddots}
% \MD{} supplies also \cs{iddots}, with behavior similar to \cs{ddots},
% except it produces the other diagonal. This command is not provided in
% basic \TeX{} or \LaTeX{}.
%
% \DescribeMacro{\vdots}
% \cs{vdots} produces vertical dots (as in standard \TeX{} or \LaTeX{}),
% but again, the \MD{} version will change size with size
% changing commands.
%
% \DescribeMacro{\fixedddots}
% \DescribeMacro{\fixedvdots}
% \DescribeMacro{\fixediddots}
% The old behavior is stored in \cs{fixedddots} and \cs{fixedvdots}.
% If you wish \cs{iddots} to behave similarly then use \cs{fixediddots}.
%
% The \cs{fixed...} versions try to maintain the (awful) default behavior,
% which is to let the horizontal dimensions change size, as well as the
% dots themselves, but hard code unchanging vertical dimensions. The
% `fixed' part of the name does not mean `corrected', but rather
% `unchanging' or `frozen'.
%
% \DescribeMacro{\originalddots}
% \DescribeMacro{\originalvdots}
% \DescribeMacro{\originaliddots}
% On the chance that some other package changes the behavior before \MD{}
% is loaded, use \cs{originalddots}, \cs{originalvdots}, and
% \cs{originaliddots} to obtain the versions in force before \MD{} is
% loaded.
%
% \DescribeMacro{\dddot}
% \DescribeMacro{\ddddot}
% If the \prog{amsmath} package is detected in \LaTeX{}, the commands
% \cs{dddot} and \cs{ddddot} are modified to get the same size changing
% behavior in 'scripts. They also correct a positioning bug in the
% \prog{amsmath} code. Compare the old and new versions and the 2-dot accent
% (from basic \LaTeX{}): $\originaldddot{X},\dddot{X},\ddot{X}$. I have
% also taken the liberty of changing the spacing between dots to match
% that in the 2-dot accent.
%
% \DescribeMacro{\originaldddot}
% \DescribeMacro{\originalddddot}
% Some of the changes to \cs{dddot} and \cs{ddddot} require the command
% \cs{text}, and so is only applied if \prog{amsmath} has been loaded
% \emph{before} \MD{}. The old amsmath commands are saved in
% \cs{originaldddot} and \cs{originalddddot}. Similar comments apply to
% plain \TeX{} with regard to \file{amstex.tex}.
%
% \DescribeMacro{\MDoprekern}
% \DescribeMacro{\MDodotkern}
% \DescribeMacro{\MDopostkern}
% Different math fonts have different metrics for the \cs{dot} and
% \cs{ddot} accents. If you use a math font family other than the CM
% fonts, you can adjust the positioning and spacing of the the dots in
% \cs{dddot} and \cs{ddddot} to better match the spacing in \cs{ddot}
% with the following length commands: \cs{MDoprekern}, \cs{MDodotkern} and
% \cs{MDopostkern}. These adjust the space before the first dot, between
% the dots and after the last dot, respectively. They are in units of
% \texttt{mu}, which is $1/3$ the size of the space made by \cs{,}.
% Reducing the first and/or increasing the last one cause the dots to be
% shifted left. Reducing the middle one causes the dots to be closer
% together. As there is usually a bit of white on either side of the dot
% character, these usually have to be given negative values (so ``reduce''
% means ``make more negative''). The defaults for the CM fonts are
% accomplished by the commands
% \begin{verbatim}
%   \MDoprekern=0mu
%   \MDodotkern=-1.3mu
%   \MDopostkern=-1mu \end{verbatim}
% I have found that the \prog{mathptmx} package (Times fonts in text and
% math) benefits from changing these via the following commands
% \begin{verbatim}
%   \MDoprekern=-1mu
%   \MDodotkern=-2mu
%   \MDopostkern=0mu \end{verbatim}
% \cs{MDoprekern}, \cs{MDodotkern}, and \cs{MDopostkern} were made
% user-accessible commands in version 0.9.
%
% \DescribeMacro{\MDoddots}
% \DescribeMacro{\MDovdots}
% \DescribeMacro{\MDoiddots}
% \DescribeMacro{\MDodddot}
% \DescribeMacro{\MDoddddot}
% In \LaTeX{}, all these commands have been made robust as of version 0.6.
% On the chance that some later loaded package changes things, the \MD{}
% versions can be obtained with \cs{MDoddots}, \cs{MDovdots},
% \cs{MDoiddots}, \cs{MDodddot} and \cs{MDoddddot} (as of version 0.9).
%
% \section{The implementation}
%
% The following are commands for plain \TeX{} that prevent double loading,
% announce the package name, and change/restore the category of \mtt{@}.
%    \begin{macrocode}
%<*tex>
\expandafter
\ifx\csname MathDots\mathdotsfileversion\endcsname\relax
\else
  \immediate\write16{Mathdots already loaded.}\expandafter\endinput
\fi
\immediate\write16{Package mathdots, \mathdotsfiledate\space version
  \mathdotsfileversion.}%
\expandafter\edef\csname MathDots\mathdotsfileversion\endcsname{%
    \catcode`\noexpand\@=\the\catcode`@}%
\catcode`\@=11
\def\@nameuse#1{\csname#1\endcsname}%
%</tex>
%    \end{macrocode}
%
% Now various shorthands. For example `\cs{MDo@us}\cs{MDo@t}' becomes
% `\cs{textstyle}' and `\cs{MDo@uf}\cs{MDo@ss}' produces
% `\cs{scriptscriptfont0}'.
%    \begin{macrocode}
\def\MDo@us#1{\@nameuse{#1style}}%
\def\MDo@uf#1{\@nameuse{#1font}0}%
\def\MDo@t{text}\def\MDo@s{script}\def\MDo@ss{scriptscript}%
\newdimen\MDo@unit \MDo@unit\p@
%    \end{macrocode}
%
% The following are the main utility macros to implement changes of style.
% Since most of our symbols are built with boxes, they have to exit math
% mode and need \cs{mathchoice} so we know the current style (i.e., size)
% when we re-enter math mode.
%
% We detect differences in sizes (for example, after \cs{large}) by
% examining the appropriate \cs{fontdimen} of family 0. Instead of the
% hard coded \mtt{1pt}, we use \mtt{0.1em} from this family.
%
% We pass along the current style via \cs{everymath}.
%
% In \cs{MDo@palette}, \arg1 is a command which we feed the current
% style name. For example, \cs{MDo@palette}\cs{MDo@ddots} will be our
% definition of \cs{ddots}. Then \cs{MDo@ddots} will see the current style
% as its first argument.
%    \begin{macrocode}
\def\MDo@changestyle#1{\relax\MDo@unit0.1\fontdimen6\MDo@uf{#1}%
  \everymath\expandafter{\the\everymath\MDo@us{#1}}}%
\def\MDo@dot{$\m@th\ldotp$}%
\def\MDo@palette#1{\mathchoice{#1\MDo@t}{#1\MDo@t}{#1\MDo@s}{#1\MDo@ss}}%
%    \end{macrocode}
%
% In the following commands, \arg1 is the current style (supplied via
% \cs{MDo@palette}). Except for the change in the unit used and the dot
% used, the first two definitions are essentially the same as plain's
% \cs{ddots} and \cs{vdots}, and the third is the same as the first with
% the boxes reversed.
%    \begin{macrocode}
\def\MDo@ddots#1{{\MDo@changestyle{#1}%
  \mkern1mu\raise7\MDo@unit\vbox{\kern7\MDo@unit\hbox{\MDo@dot}}%
  \mkern2mu\raise4\MDo@unit\hbox{\MDo@dot}%
  \mkern2mu\raise \MDo@unit\hbox{\MDo@dot}\mkern1mu}}%
\def\MDo@vdots#1{\vbox{\MDo@changestyle{#1}%
  \baselineskip4\MDo@unit\lineskiplimit\z@
  \kern6\MDo@unit\hbox{\MDo@dot}\hbox{\MDo@dot}\hbox{\MDo@dot}}}%
\def\MDo@iddots#1{{\MDo@changestyle{#1}%
  \mkern1mu\raise \MDo@unit\hbox{\MDo@dot}%
  \mkern2mu\raise4\MDo@unit\hbox{\MDo@dot}%
  \mkern2mu\raise7\MDo@unit\vbox{\kern7\MDo@unit\hbox{\MDo@dot}}}}%
%    \end{macrocode}
%
% \DescribeMacro{\fixedddots}
% \DescribeMacro{\fixedvdots}
% We include current (as of 2014/06/11) \LaTeX{}/plain\TeX{} definitions
% so a user can choose them. \LaTeX{} and plain\TeX{} don't have the other
% diagonal, so we provide a \DescribeMacro{fixediddots}\cs{fixediddots}
% which is just \cs{fixedddots} with the boxes reversed.
%
% \DescribeMacro{\originalvdots}
% \DescribeMacro{\originalddots}
% We save the versions at load time so user can use them in case he
% loads \MD{} for its other features. Chances are the original
% \DescribeMacro{\originaliddots}\cs{iddots} is undefined.
%    \begin{macrocode}
%<sty>\DeclareRobustCommand\fixedvdots{%
%<tex>\def\fixedvdots{%
  \vbox{\baselineskip4\p@ \lineskiplimit\z@
  \kern6\p@\hbox{.}\hbox{.}\hbox{.}}}%
%<sty>\DeclareRobustCommand\fixedddots{%
%<tex>\def\fixedddots{%
  \mathinner{\mkern1mu
  \raise7\p@\vbox{\kern7\p@\hbox{.}}\mkern2mu
  \raise4\p@\hbox{.}\mkern2mu
  \raise\p@\hbox{.}\mkern1mu}}%
%<sty>\DeclareRobustCommand\fixediddots{%
%<tex>\def\fixediddots{%
  \mathinner{\mkern1mu
  \raise\p@\hbox{.}\mkern2mu
  \raise4\p@\hbox{.}\mkern2mu
  \raise7\p@\vbox{\kern7\p@\hbox{.}}\mkern1mu}}%
\let\originalddots=\ddots
\let\originalvdots=\vdots
\let\originaliddots=\iddots
%    \end{macrocode}
%
% \DescribeMacro{\ddots}
% \DescribeMacro{\vdots}
% \DescribeMacro{\iddots}
% Here are the actual (re)definitions of these three commands. The
% plain\TeX{} version emits messages similar to that produced by
% \LaTeX{}'s \cs{DeclareRobustCommand}.
%
% Since \cs{ddots} et al.\ are defined in \LaTeX{} with
% \cs{DeclareRobustCommand}, saving the original definition with
% \verb$\let\originalddots\ddots$ will not work unless the internal
% command ``\verb*$\ddots $'' is unchanged. Thus we can't use
% \verb$\DeclareRobustCommand\ddots$ directly because it changes that
% internal. Instead we define \cs{MDoddots} and then do
% \verb$\let\ddots\MDoddots$
%    \begin{macrocode}
%<sty>\DeclareRobustCommand\MDoddots{%
%<tex>\def\MDoddots{%
  \mathinner{\MDo@palette\MDo@ddots}}%
%<sty>\DeclareRobustCommand\MDoiddots{%
%<tex>\def\MDoiddots{%
  \mathinner{\MDo@palette\MDo@iddots}}%
%<sty>\DeclareRobustCommand\MDovdots{%
%<tex>\def\MDovdots{%
  \mathinner{\MDo@palette\MDo@vdots}}%
%<tex>\wlog{Mathdots Info: Redefining \string\ddots.}%
%<tex>\wlog{Mathdots Info: Redefining \string\vdots.}%
\let\ddots \MDoddots
\let\iddots\MDoiddots
\let\vdots \MDovdots
%    \end{macrocode}
%
% Now we try to get AMS \cs{dddot} and \cs{ddddot} accents to behave
% nicely.
%
% \DescribeMacro{\dddot}
% \DescribeMacro{\ddddot}
% We add a \verb"\kern\z@" to prevent \TeX{} from vertically centering
% \arg1. I don't know if the \cs{text} command is necessary, but it
% doesn't seem to hurt. I included it in the original version because that
% one used text dots and needed the text to change size. Now I use a math
% symbol (the dot accent) and \cs{mathpalette} to implement the change.
% Still, I don't really understand how \cs{ex@} works in the AMS packages,
% so I leave it in.
% \DescribeMacro{MDoprekern}
% \DescribeMacro{MDodotkern}
% \DescribeMacro{MDopostkern}
% For possible fine tuning, I have used two macros that
% represent the math spacing placed before and after the sequence of
% dots (\cs{MDoprekern} and \cs{MDopostkern}). Also, the actual dot is in
% the definition of \cs{MDo@D} where another spacing command follows the
% dot (\cs{MDodotkern}). In fact, by redefining \cs{MDo@D} and changing
% the spacing, one can define other multiple ``dot'' accent commands.
%
% \DescribeMacro{\originaldddot}
% \DescribeMacro{\originalddddot}
% We also save the original \prog{amsmath} versions.
%    \begin{macrocode}
\newbox\MDo@dotsbox
\newmuskip\MDoprekern
\newmuskip\MDopostkern
\newmuskip\MDodotkern
\MDoprekern 0mu
\MDopostkern -1mu
\MDodotkern -1.3 mu
\def\MDo@dotsaccent#1{\mathpalette{\MDo@@dotsaccent{#1}}}%
\let\originaldddot \dddot
\let\originalddddot\ddddot
%<*sty>
\def\MDo@@dotsaccent#1#2#3{%
  {\setbox\MDo@dotsbox\hbox{$#2\mkern\MDoprekern#1\mkern\MDopostkern$}%
   \mathop{#3\kern\z@}\limits^{%
   \text{\vbox to-1.4\ex@{\kern-1.8\ex@\copy\MDo@dotsbox\vss}}}}%
}%
\@ifpackageloaded{amsmath}{%
  \def\MDo@D{\mathchar"5F\mkern\MDodotkern}%
  \DeclareRobustCommand\MDodddot {%
    \MDo@dotsaccent{\MDo@D\MDo@D\MDo@D}}%
  \DeclareRobustCommand\MDoddddot{%
    \MDo@dotsaccent{\MDo@D\MDo@D\MDo@D\MDo@D}}%
  \let\dddot \MDodddot
  \let\ddddot\MDoddddot
}{}%
%</sty>
%    \end{macrocode}
%
% For plain tex + amstex: \cs{ex@} is set in \file{amstex.tex} to
% \texttt{.2326ex}, so its value depends on the value of \texttt{ex} when
% \prog{amstex} is loaded. Unlike \prog{amsmath}, it doesn't seem to be
% recalculated with size changes. Thus, in plain TeX we add a reset.
%    \begin{macrocode}
%<*tex>
\def\MDo@@dotsaccent#1#2#3{%
{%
  \setbox\MDo@dotsbox\hbox{$#2\mkern\MDoprekern#1\mkern\MDopostkern$}%
  \mathop{#3\kern\z@}\limits^{\text{\ex@.2326ex
    \vbox to-1.4\ex@{\kern-1.8\ex@\copy\MDo@dotsbox\vss}}}%
}}%
\ifx\amstexloaded@\relax % true if amstex has been loaded
  % A backup test in case \amstexloaded@ became \relax by accident:
  \ifx\ddddot\UndEfInEd
  \else
    \def\MDo@D{\mathchar"5F\mkern\MDodotkern}%
    \wlog{Mathdots Info: Redefining \string\dddot.}%
    \wlog{Mathdots Info: Redefining \string\ddddot.}%
    \def\MDodddot {\MDo@dotsaccent{\MDo@D\MDo@D\MDo@D}}%
    \def\MDoddddot{\MDo@dotsaccent{\MDo@D\MDo@D\MDo@D\MDo@D}}%
    \let\dddot \MDodddot
    \let\ddddot\MDoddddot
  \fi
\fi
\csname MathDots\mathdotsfileversion\endcsname
%</tex>
%</sty|tex>
%    \end{macrocode}
%\Finale