% \iffalse
% File: examdesign.dtx Copyright (C) 1997-2001 Jason Alexander
% Copyright (C) 19yy  name of author
%
% This program is free software; you can redistribute it and/or
% modify it under the terms of the GNU General Public License
% as published by the Free Software Foundation; either version 2
% of the License, or (at your option) any later version.
%
% This program is distributed in the hope that it will be useful,
% but WITHOUT ANY WARRANTY; without even the implied warranty of
% MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
% GNU General Public License for more details.
%
% You should have received a copy of the GNU General Public License along with
% this program; if not, write to the Free Software Foundation, Inc., 59 Temple
% Place - Suite 330, Boston, MA  02111-1307, USA.
%
% For error reports, or offers to help improve the 'examdesign' class 
% please contact me at jalex@uci.edu.
% \fi
%         \ProvidesFile{examdesign.dtx}
%          [1997/11/22 v1.02 `examdesign' class (JMA)]
% \iffalse
%<*driver>
\documentclass{ltxdoc}
\usepackage{tabularx}
\CodelineIndex
\EnableCrossrefs
\RecordChanges
\begin{document}
\DocInput{examdesign.dtx}% \PrintIndex
\end{document}
%</driver>
%\fi
% \MakeShortVerb{\|}
% \MakeShortVerb{\"}
%
% \def\fileversion{v1.1}
% \def\filedate{2001/3/26}
% \changes{v1.00}{1997/10/11}{Initial version.}
% \changes{v1.01}{1997/10/22}{Fixed several bugs, added
%      \texttt{StudentInfoLineSpacing}
%      command, included two example files
%      to better illustrate the class features.}
% \changes{v1.02}{1997/11/22}{Added the \textsf{block}, \textsf{frontmatter},
%       \textsf{endmatter}, \textsf{examtop}, \textsf{keytop}, \textsf{exampreface},
%       \textsf{keypreface}, \textsf{examclosing}, \textsf{keyclosing} environments
%       and the commands \texttt{word}, \texttt{exam}, \texttt{key},
%       \texttt{examvspace}, \texttt{examvspace*}, \texttt{examhspace},
%       \texttt{examhspace*}, \texttt{keyvspace}, \texttt{keyvspace*},
%       \texttt{keyhspace}, and \texttt{keyhspace*}. Fixed a bug that prevented
%       the \textsf{multiplechoice} environment from identifying the answer
%       in the answer key if the correct answer was the last choice.  Added
%       the \texttt{onekey} option. Fixed a careless bug that broke the
%       cedilla.}
% \changes{v1.1}{2001/3/26}{Reworked the internal code, added the
% \textsf{matching} environment, and changed the input syntax considerably to
% make it (I hope) more user-friendly. Questions can now be numbered
% continously across sections, plus more\ldots}
% \changes{v1.101}{2006/09/05}{Jim Hefferon Changed name of files and class to
% examdesign to avoid name clashes with another package, after getting OK
% from author.}
%
% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%
% \title{The \textsf{examdesign} class\thanks{This file
%        has version number \fileversion, last
%        revised \filedate.}}
% \author{Jason Alexander}
% \date{\filedate}
% \maketitle
% \hfuzz=72pt
% \vfuzz=36pt
%
% \DoNotIndex{\if,\else,\fi,\expandafter}
% \DoNotIndex{\csname,\endcsname}
% \DoNotIndex{\addtocounter,\stepcounter,\advance,\Alph,\alph,\arabic}
% \DoNotIndex{\begin,\end,\begingroup,\endgroup,\catcode,\centerline,\day,\def}
% \DoNotIndex{\divide,\do,\edef,\endgroup,\evensidemargin,\fbox,\gdef,\global}
% \DoNotIndex{\headheight,\headsep,\hfill,\hphantom,\Huge,\ifnum,\ifx,\large}
% \DoNotIndex{\long,\m@ne,\mbox,\medskip,\message,\month,\multiply}
% \DoNotIndex{\noindent,\nopagebreak,\normalsize,\oddsidemargin,\pagebreak}
% \DoNotIndex{\par,\relax,\rightmargin,\roman,\setcounter,\setlength}
% \DoNotIndex{\smallskipamount,\space,\textheight,\textsf,\textwidth}
% \DoNotIndex{\the,\topmargin,\undefined,\underline,\value,\vspace}
% \DoNotIndex{\xdef,\year,\z@,\time,\topsep}
%
% \begin{abstract}
% A new class, \textsf{examdesign}, for \LaTeXe\ is defined.  It provides
% several features useful for designing tests or question sets: (1) it allows
% for an explicit markup of questions and answers; (2) the class will, at the
% user's request, automatically generate answer keys; (3) multiple versions of
% the same test can be generated automatically, with the ordering of questions
% within each section randomly permuted so as to minimize cheating; (4) the
% generated answer keys can be constructed either with or without the
% questions included; and (4) some environments are provided to assist in
% constructing the most common types of test question: matching,
% true/false, multiple-choice, fill-in-the-blank, and short answer/essay
% questions
% \end{abstract}
%
% \section{Description}
%
% Teaching is a rewarding and enjoyable profession, designing exams is not.
% \emph{Grading} exams is even worse, but for help with that last bit you'll
% need more than a \LaTeX\ class.  This class file started out with a
% relatively simply goal: randomize a collection of questions so that the
% author wouldn't need to do it by hand.  Since then, it has grown
% in functionality and size and is now (I hope) a reasonably complete
% general-purpose class for constructing exams and question sets.
%
%
% \section{Usage}
%
% To begin, simply create a new \LaTeXe\ document starting with
% \begin{verbatim}
% \documentclass{examdesign}
% \begin{document}\end{verbatim}
% If you want to create an exam in 12pt type, use
% "\documentclass[12pt]{examdesign}" instead.  If you need to load
% additional packages, they may be loaded via the command "\usepackage" as
% always.  The class includes a number of packages included with a standard \LaTeX\
% distribution: \textsf{keyval}, \textsf{multicol}, and \textsf{enumerate}.
% If you don't have these packages, your \LaTeX\ distribution is nonstandard
% (or incomplete), and you'll have to grab these packages from the nearest
% CTAN site in order to use \textsf{examdesign}.
%
% Most exams/questions sets are split into sections according to the type of
% question asked, and \textsf{examdesign} follows this practice.  There are five
% environments defined, one environment for each type of question: \textsf{matching},
% \textsf{shortanswer}, \textsf{truefalse}, \textsf{multiplechoice}, and
% \textsf{fillin}.  \textbf{N.B.} Previous versions of \textsf{examdesign} had
% starred forms of these environments of well.  The starred forms (as well as
% the \textsf{fixed} environment, have been eliminated in this new release.
% Their functionality has been incorporated into the standard environments.
%
% All the environments are designed to make binding answers to questions
% easier. This not only makes the job of writing the exam easier, 
% but it allows the automated generation of answer keys.
%
% \subsection{Switches}
% There are two ways to customize the output of the \textsf{exam}
% class: switches and special formatting environments.
% Descriptions of the special formatting environments (such as
% \textsf{frontmatter}, \textsf{endmatter}, etc.) can be found in section
% \ref{SFE}. 
%
% A switch is a single command that changes an internal parameter
% influencing the overall formatting of the exam.  Switches should be included
% in the preamble because doing so will guarantee that they will apply to the
% entire exam.  Putting
% switches elsewhere may result in odd results because \textsf{examdesign} shuffles
% around what you write quite a lot, and you won't always know when a
% switch will be executed.
%
% One change between this release and previous releases is that all of the
% class options unique to \textsf{examdesign} have been converted to switches for
% sake of consistency. 
%
% \DescribeMacro{\NoRearrange}
% Although the original reason behind writing \textsf{examdesign} was to automate
% the random rearranging of questions, people have pointed out that it would
% be nice to turn off the feature (say, when writing the exam).  The
% "\NoRearrange" switch suppresses the rearrangement of questions.  You can
% override this setting on a section-by-section basis, though, by setting the
% correct environment option (see the discussion of the options in
% section~\ref{env} for more details).  In addition, see the discussion of the
% "\setrandomseed" macro below for how to generate ``randomly rearranged''
% question sets in the \emph{same} order across runs.
%
% \DescribeMacro{\NoKey}
% \DescribeMacro{\OneKey}
% The switch "\NoKey" suppresses the generation of all answer keys, whereas
% "\OneKey" generates an answer key for the first exam only.  
%
% \DescribeMacro{\BoldfaceCorrectMultipleChoiceAnswer}
% \DescribeMacro{\UnderlineCorrectMultipleChoiceAnswer}
% The switch "\BoldfaceCorrectMultipleChoiceAnswer" identifies the correct
% answer in multiple choice questions by using boldface text instead of
% placing a box around the correct letter.  The switch
% "\UnderlineCorrectMultipleChoiceAnswer" identifies the correct answer in
% multiple 
% choice questions by underlining.  These two switches are cumulative:
% including both in your exam will 
% give you boldface underlined answers (the order doesn't matter).  If neither
% one is specified, 
% \textsf{examdesign} will default to boxing the letter of the correct answer.  This
% is shown in the examples below. 
%
% Technical note: The ordinary \LaTeXe\ "\underline" command does not work
% well when the underlined text spans several lines. Consequently,
% I've incorporated the \textsf{ulem} package, written
% by Donald Arseneau, for underlining.  Since this package is embedded
% in this class, you do not need to load it again if you should need it.
%
% \DescribeMacro{\NumberOfVersions}
% The "\NumberOfVersions" macro takes a single numerical argument---the number of
% different versions of the test to generate.  Under the default
% settings, if you request more than 26 versions of the test you will
% get an error.  This is because the forms are labelled according to letters of
% the alphabet, and \LaTeXe\ doesn't like it when you try to ask
% for the 27th letter of the alphabet.  If you \emph{really} need more than 27
% copies of an exam, you will need to use one of the
% special formatting environments redefine the material inserted at the top of
% each exam and key.
%
% \DescribeMacro{\class}
% The "\class" macro takes one argument: the name of the class to appear on
% the exam.  For example, "\class{Math 205A: Set Theory}" will cause the
% name of the class to appear in the appropriate place (if you are using the
% default setting).
%
% \DescribeMacro{\examname}
% |\examname| takes one argument, just like |\class|.  The argument is used to
% identify the type of test.  For example, |\examname{Final Exam}|,
% |\examname{Midterm Exam}|, |\examname{Pop Quiz}|,\ldots
%
% \DescribeMacro{\ConstantBlanks}
% |\ConstantBlanks| has one mandatory argument (a length).  It tells \TeX\ to
% typeset the underline in fill-in-the-blank environments using lines of a
% constant length, regardless of the length of the correct word.
%
% \DescribeMacro{\ProportionalBlanks}
% |\ProportionalBlanks| has one mandatory argument (a real number $r$). It tells \TeX\
% to typeset the underline in fill-in-the-blank environments using lines that
% are exactly $r$ times the length of the correct answer when the correct
% answer is typeset by \TeX.  The default setting is |\ProportionalBlanks{1}|.
% If you include both "\ConstantBlanks" and "\ProportionalBlanks" in the preamble,
% the exam will be typeset using whichever switch was last set.
%
% \DescribeMacro{\StudentInfoLineSpacing}
% |\StudentInfoLineSpacing| requires a length as its argument.  This value
% is used to determine the amount of additional spacing between lines in
% the student data box appearing in the upper right-hand corner of the exam.
% The default value is 6pt.  If you change the definitions these environments
% using \textsf{examtop} and \textsf{keytop}, this command will have no effect.
%
% \DescribeMacro{\SectionFont}
% This macro takes one argument, which can be any declarative font-changing
% commands (i.e., "\em" rather than "\emph").  It changes the default font
% used to typeset the section titles.
%
% \DescribeMacro{\ContinuousNumbering}
% This macro takes no argument.  It changes the default numbering behavior.
% Instead of resetting the numbers at the start of each section, it will
% continue the numbering from previous sections.  (Notice that the
% \textsf{matching} and \textsf{truefalse} environments are not affected by
% this, as they don't number their questions.)
%
% \DescribeMacro{\ShortKey}
% This macro takes no argument.  It changes the way answer keys are typeset.
% Basically, it will delete any question text, any instructions at the start
% of a block of questions (or at the start of a section), and only print the
% answer.  
%
% \DescribeMacro{\DefineAnswerWrapper}
% This macro takes two arguments.  For the \textsf{shortanswer} environment,
% the first argument will always be inserted before the answer, and the second
% argument will always be inserted after the answer.  
%
% \DescribeMacro{\SectionPrefix}
% This macro takes one argument, which specifies the default appearance of the
% section number preceding each title.  Its default setting is\hfil\break
% "\SectionPrefix{Section \arabic{sectionindex}. \space}"
%
%
% \subsubsection{Summary of Switches}
%
% \medskip
%
% \begin{tabular}{p{2in}l}
% {\large Switch Name} \rule[-3pt]{0pt}{15pt} & {\large Function} \\ \hline
% "\NoKey" &  Suppress all answer keys \\
% "\OneKey" & Generate only the first answer key, suppressing all others \\
% \multicolumn{2}{l}{\texttt \textbackslash BoldfaceCorrectMultipleChoiceAnswer See above}\\
% \multicolumn{2}{l}{\texttt \textbackslash UnderlineCorrectMultipleChoiceAnswer See above}\\
% "\NumberOfVersions" & specify number of exams to make \\ 
% "\class" & Class data (E.g., ``Philosophy 29'') \\ 
% "\examname" & Exam name (E.g., ``Midterm Exam'') \\ 
% "\ConstantBlanks" & See above \\ 
% "\ProportionalBlanks" & See above \\ 
% "\StudentInfoLineSpacing" & Set spacing in data box \\
% "\SectionFont" & Change section font\\
% "\ContinuousNumbering" & Change numbering style\\
% "\ShortKey" & Change key formatting\\
% "\DefineAnswerWrapper" & Change tags surrounding answers\\
% "\SectionPrefix" &  Change the appearance of the section number\\
% \end{tabular}
% \MakeShortVerb{\|}
%
% \subsection{Environments\label{env}}
%
% There has been a significant change in the environments between this version
% of \textsf{examdesign} and earlier versions.  In this version, all question
% environments can take an optional single argument which can be used to set a
% number of different parameters.  (Before, the optional argument was used to
% set the title for a section.)  Exams written for the previous version will
% have to be updated to work with the new version.  Sorry.
%
% The options available for each section are:
%
% \noindent
% \begin{tabularx}{\linewidth}{>{\tt}lX}
% Name & Function\\\hline
% title & Specifies the title for each section.  When no title is specified,
% no section number is displayed (giving an entirely blank header before the
% section). However, the section instructions will still be printed.  If you
% would like a section number \emph{without} a title, simply give a blank
% space for the title, or use "\relax", as in "title={ }" or "title={\relax}".\\
% rearrange & This option takes one of two values, either "yes" or "no".  If
% "yes",then the questions in that section are randomly rearranged; if "no",
% then questions in that section appear in the order in which they are
% written.  This option works for all environments (and hence replaces the old
% \textsf{fixed} environment).\\
% resetcounter & This option takes one of two values, either "yes" or "no".
% If "yes", then the question numbers for that section start at 1, overriding
% the "\ContinuousNumbering" switch.  If "no", then the question numbers for
% that section will not be reset.  This gives you section-level control over
% the numbering.\\
% keycolumns & This option should be set to a number.  It specifies the number
% of columns that should be used when typesetting the section in the answer
% key.  For example, if you use the "\ShortKey" switch to only include the
% correct answer for multiple-choice questions, each answer takes up very
% little space on the line.  If you have 40 questions, it would take 40
% lines---a waste of paper.  Setting "keycolumns=5" would typeset the answer
% key using 5 columns, requiring only 8 lines for the answers.  The section
% instructions will \emph{not} be typeset in columns, but will span the width
% of the page, instead.  Note that it is very easy to get overfull/underfull
% box messages with this option (and the next).\\
% examcolumns & Same as above, except it affects the exam.\\
% suppressprefix & Can be set to either "yes" or "no".  If "yes", it
% suppresses the section prefix for that particular section; if "no", the
% prefix will appear as always.  This option replaces the starred environments
% in the the previous release.\\
%\end{tabularx}
%
% \subsubsection{The \textsf{fillin} environment}
%
% This environment is for fill-in-the-blank questions.  
% The syntax is:
% \begin{verbatim}
% \begin{fillin}[title={Insert title here}]
%   \begin{question}
%     How much \blank{wood} could a \blank{woodchuck} chuck if a woodchuck could
%     \blank{chuck} wood?
%   \end{question}
% \end{fillin}\end{verbatim}
% When typeset using "\ProportionalBlanks", the questions will be printed in
% the exam as: 
% \def\blank#1{$\underline{\mbox{\hphantom{#1}}}$}
% \begin{enumerate}
%   \item How much \blank{wood} could a \blank{woodchuck} chuck if a woodchuck could
%         \blank{chuck} wood?
% \end{enumerate}
% but in the answer key as:
% \def\blank#1{$\underline{\mbox{#1}}$}
% \begin{enumerate}
%   \item How much \blank{wood} could a \blank{woodchuck} chuck if a woodchuck could
%         \blank{chuck} wood?
% \end{enumerate}
%
% \subsubsection{The \textsf{shortanswer} environment}
%
% The \textsf{shortanswer} environment does not format its questions in any
% way other than enumerating them.  (It is then a catch-all environment for
% any other type of question.)  It does allow the group of
% answers with questions. Including an answer is not mandatory (but you will get
% errors if you do not include an answer and ask for a "\ShortKey").
% The general syntax is:
% \begin{verbatim}
% \begin{shortanswer}[title={Interesting questions...},
%                     rearrange=yes]
%   \begin{question}
%     State Hobbes' definition of the state of nature and the role it plays in
%     his social philosophy.
%     \begin{answer}
%       The state of nature is\ldots
%     \end{answer}
%   \end{question}
%
%   \begin{question}
%     State several examples that seem to illustrate that Mill's theory of
%     Utilitarianism cannot be followed in practice, and explain why.
%     Afterwards, state Mill's response to the problems you raised.
%     \begin{answer}
%       Often times it appears we need to act on our moral instincts without
%       taking time to deliberate (as in the case of saving a drowning child).
%       According to Mill, though, it seems that before each moral action we
%       ought to deliberate in order to be sure that the action we take is such
%       as to maximize the overall general utility\ldots
%     \end{answer}
%   \end{question}
% \end{shortanswer}
% \end{verbatim}
% \vspace{-\baselineskip}
% Note that the \textsf{answer} environment appears inside the \textsf{question}
% environment.  
%
% In the exam, this will be typeset as:
% \begin{enumerate}%
%   \item State Hobbes' definition of the state of nature and the role it plays in
%         his social philosophy.
%   \item State several examples that seem to illustrate that Mill's theory of
%     Utilitarianism cannot be followed in practice, and explain why.
%     Afterwards, state Mill's response to the problems you raised.
% \end{enumerate}%
% The answer key, in normal form with the default definition of
% "\DefineAnswerWrapper", will appear as: 
% \begin{enumerate}%
%   \item State Hobbes' definition of the state of nature and the role it plays in
%         his social philosophy.
%         \begin{description}%
%            \item[Answer:] The state of nature is\ldots
%         \end{description}%
%   \item State several examples that seem to illustrate that Mill's theory of
%         Utilitarianism cannot be followed in practice, and explain why.
%         Afterwards, state Mill's response to the problems you raised.
%         \begin{description}%
%            \item[Answer:] Often times it appears we need to act on our moral
%                instincts without taking time to deliberate (as in the case of
%                saving a drowning child). According to Mill, though, it seems that
%                before each moral action we ought to deliberate in order to be sure
%                that the action we take is such as to maximize the overall general
%                utility\ldots
%         \end{description}%
% \end{enumerate}%
% If you specify a "\ShortKey", only the answers will be presented:
% \begin{enumerate}%
%   \item \begin{description}
%            \item[Answer:] The state of nature is\ldots
%         \end{description}
%   \item \begin{description}
%            \item[Answer:] Often times it appears we need to act on our moral
%                instincts without taking time to deliberate (as in the case of
%                saving a drowning child). According to Mill, though, it seems that
%                before each moral action we ought to deliberate in order to be sure
%                that the action we take is such as to maximize the overall general
%                utility\ldots
%         \end{description}
% \end{enumerate}
%
%
% \subsubsection{The \textsf{truefalse} environment}
%
% True/false questions.  The general syntax is:
% \begin{verbatim}
% \begin{truefalse}
%   \begin{question}
%     \answer{False} Laden swallows fly faster than unladen swallows,
%     especially if they carry coconuts.
%   \end{question}
%
%   \begin{question}
%     \answer{True} Quantum mechanics was first stated using the
%     mathematical tools of Hilbert spaces by John von~Neumann.
%   \end{question}
% \end{truefalse}\end{verbatim}
%
% Again, note that the \textsf{answer} environment appears within the
% \textsf{question} environment.  The above will be typeset in the exam as:
% \begin{list}{$\underline{\hskip.5in}$}{\setlength{\rightmargin}{0pt}
%                \setlength{\labelwidth}{.5in}
%                \setlength{\leftmargin}{.6in}
%                \setlength{\labelsep}{.1in}
%                \setlength{\itemindent}{0in}
%                \setlength{\topsep}{4pt}
%                \setlength{\itemsep}{.1in}}
%   \item Laden swallows fly faster than unladen swallows, especially if they
%     carry coconuts.
%   \item Quantum mechanics was first stated using the mathematical tools of Hilbert
%     spaces by John von~Neumann.
% \end{list}
% It will be typeset in the answer key as:
% \begin{list}{$\underline{\hphantom{False}}$}{\setlength{\rightmargin}{0pt}
%                \setlength{\labelwidth}{.5in}
%                \setlength{\leftmargin}{.6in}
%                \setlength{\labelsep}{.1in}
%                \setlength{\itemindent}{0in}
%                \setlength{\topsep}{4pt}
%                \setlength{\itemsep}{.1in}}
%   \item[$\underline{\makebox[.5in][c]{False}}$] Laden swallows fly faster than unladen
%     swallows, especially if they carry coconuts.
%   \item[$\underline{\makebox[.5in][c]{True}}$] Quantum mechanics was first
%   stated using the mathematical tools of Hilbert spaces by John von~Neumann.
% \end{list}
%
% \subsubsection{The \textsf{multiplechoice} Environment}
%
% It isn't difficult to format multiple choice questions in ordinary \LaTeX\ with
% multiply nested \textsf{enumerate}'s, but this environment provides for a
% simpler entry scheme, and the answer key can indicate the correct answer.
% The syntax is: 
% \begin{verbatim}
% \begin{multiplechoice}
%   \begin{question}
%     How much wood could a woodchuck chuck if a woodchuck could chuck wood?
%      \choice{A lot.}
%      \choice{More than most.}
%      \choice{Exactly $\pi$ cords.}
%      \choice[!]{It depends on the nature of the woodchuck.}
%   \end{question}
% \end{multiplechoice}
% \end{verbatim}
% \vspace{-\baselineskip}
% In the exam, this is typeset as:
% \begin{enumerate}
%  \item How much wood could a woodchuck chuck if a woodchuck could chuck wood?
% \begin{list}{$\underline{\hphantom{False}}$}{\setlength{\rightmargin}{0pt}
%                \setlength{\labelwidth}{.4in}
%                \setlength{\leftmargin}{.5in}
%                \setlength{\labelsep}{.1in}
%                \setlength{\itemindent}{0in}
%                \setlength{\topsep}{0pt}
%                \setlength{\itemsep}{0pt}}
% \item[(a)] A lot.
% \item[(b)] More than most.
% \item[(c)] Exactly $\pi$ cords.
% \item[(d)] It depends on the nature of the woodchuck.
% \end{list}
% \end{enumerate}
% In the answer key, if neither of the options |mcbold| nor |mcunderline|
% were given, the answer will be typeset as:
% \begin{enumerate}
%  \item How much wood could a woodchuck chuck if a woodchuck could chuck wood?
% \begin{list}{$\underline{\hphantom{False}}$}{\setlength{\rightmargin}{0pt}
%                \setlength{\labelwidth}{.4in}
%                \setlength{\leftmargin}{.5in}
%                \setlength{\labelsep}{.1in}
%                \setlength{\itemindent}{0in}
%                \setlength{\topsep}{0pt}
%                \setlength{\itemsep}{0pt}}
% \item[(a)] A lot.
% \item[(b)] More than most.
% \item[(c)] Exactly $\pi$ cords.
% \item[\fbox{(d)}\hspace{-\fboxsep}] It depends on the nature of the woodchuck.
% \end{list}
% \end{enumerate}
% You can have up to 26 choices. Also, notice that the correct answer was
% specified by giving the optional argument "[!]" to |\choice|. 
%
% \subsubsection{The \textsf{matching} environment}
%
% This environment is new in this release of \textsf{examdesign}.  It provides an
% environment for the creation of matching tests.  The syntax is:
% \begin{verbatim}
% \begin{matching}[title={Some matching questions}]
%   \pair{John Steinbeck}{\emph{The Grapes of Wrath}}
%   \pair{Will Self}{\emph{My Kind of Fun}}
%   \pair{Charles Darwin}{\emph{The Origin of Species}}
% \end{matching}\end{verbatim}
% In the exam, this will be typeset as:
% \begin{multicols}{2}
% \begin{quote}
% \begin{enumerate}
% \item[\underline{\hskip .5in}] Will Self
% \item[\underline{\hskip .5in}] Charles Darwin
% \item[\underline{\hskip .5in}] John Steinbeck
% \columnbreak
% \item[(a)] \emph{The Origin of Species}
% \item[(b)] \emph{The Grapes of Wrath}
% \item[(c)] \emph{My Kind of Fun}
% \end{enumerate}
% \end{quote}
% \end{multicols}
% In the answer key, this will be typeset as:
% \begin{multicols}{2}
% \begin{quote}
% \begin{enumerate}
% \item[\underline{\makebox[.5in][c]{(c)}}] Will Self
% \item[\underline{\makebox[.5in][c]{(a)}}] Charles Darwin
% \item[\underline{\makebox[.5in][c]{(b)}}] John Steinbeck
% \columnbreak
% \item[(a)] \emph{The Origin of Species}
% \item[(b)] \emph{The Grapes of Wrath}
% \item[(c)] \emph{My Kind of Fun}
% \end{enumerate}
% \end{quote}
% \end{multicols}
% The "\pair" macro preceeds the \emph{matched} pair of statements.  The
% environment will randomly shuffle the first and second columns in order to
% create the (hopefully randomized) lists.  Because of the way the randomizer
% is implemented, this environment will produce better results (i.e., more
% scrambled) results the longer the list of questions is.
%
% \textbf{N.B.} The \textsf{block} environment will break badly in this
% environment, so don't use it.
%
%
% \subsubsection{The \textsf{block} environment}
%
% The \textsf{block} environment lets you identify a group of questions,
% as well as some preceeding text, as a ``block'' that should be kept together
% even though the rest of the section gets rearranged.  (The questions inside
% the block will not be rearranged.  That's on my to-do list, but it will take
% more time than I have, and it's already been a long time since the last
% update.)  
%
% The syntax is straightforward. For example:
% \begin{verbatim}
% \begin{block}
% Here is a spot where you can put some instructions, a graph, or whatever
% you want to preceed the questions in this block.
%  \begin{question}
%    This is the first question.
%    \begin{answer}
%      This is the answer to the first question.
%    \end{answer}
%  \end{question}
%
%  \begin{question}
%    This is the second question.
%    \begin{answer}
%      This is the answer to the second question.
%    \end{answer}
%  \end{question}
%
%  \begin{question}
%    This is the third question.
%    \begin{answer}
%      This is the answer to the third question.
%    \end{answer}
%  \end{question}
% \end{block}\end{verbatim}
% Will keep those three questions together when the rearranging happens.
%
% The \textsf{block} environment may be used in \emph{any} of the previously
% described question environments, except for the \textsf{matching}
% environment.  (There were some problems with using blocks inside some of the
% other environments with earlier versions---I believe those bugs have been
% fixed.)
%
% The \textsf{block} environment can take a special optional argument,
% "questions", that specifies the number of questions in the block.
% Specifying this option sets two counters, "first" and "last", so that you
% can refer to the particular questions that fall within the block in the
% instructions.  So, for example:
% \begin{verbatim}
% \begin{block}[questions=3]
% \noindent For questions \thefirst--\thelast, consider the following data...\end{verbatim} 
% would, if the block appeared after question 13, be typeset as:
% 
% \bigskip
% \noindent For questions 14--16, consider the following data...\par\bigskip
%
% In previous versions of \textsf{examdesign}, blocks were typeset indented from the
% left margin.  This was undesireable because the block instructions could
% be misread as belonging to the end of the preceeding question.  As of this
% version, block instructions are typeset using no such indent.
%
%
% \section{Special Formatting Tools}
%
% The following environments, macros, and length parameters allow the user to
% customize virtually every aspect of the exam.  As I describe below, the default
% settings of the \textsf{examdesign} are defined using these environments.
%
% \subsection{Special Formatting Environments \label{SFE}}
%
% \DescribeEnv{frontmatter}
% Any material enclosed in the \textsf{frontmatter} environment will be
% typeset on a page (or pages) by itself before the exam (and the key, too).
% For example, if one
% wanted to make a cover sheet for the exam with the class name and a place for
% the student to write their name, i.d. number, and so forth.
%
% \DescribeEnv{endmatter}
% Any material enclosed in the \textsf{endmatter} environment will be
% typeset on a page (or pages) by itself after the exam (and the key, too).
% For example, if one wanted to place a page with several important equations,
% constants, and tables at the end of the exam for students to reference, if
% needed.
%
% \DescribeEnv{examtop}
% Any material enclosed in the \textsf{examtop} environment will be
% typeset on the next page following the \textsf{frontmatter} text, if an exam
% is being created, and will be omitted if an answer key is being created.  The
% default settings of the \textsf{examdesign} class use this environment to create
% the area for the class name, student name, etc. as follows:
% \begin{verbatim}
% \begin{examtop}
%   \@@line{\parbox{3in}{\classdata \\
%           \examtype, Form: \fbox{\textsf{\Alph{version}}}}
%           \hfill
%           \parbox{3in}{\normalsize \namedata}}
%           \bigskip
% \end{examtop}\end{verbatim}
% Where "\classdata" and "\examtype" are macros that use
% the current values assigned by the switches "\class" and "\examname".
% 
%
% The definition of "\namedata" is:
% \begin{verbatim}
% \def\namedata{Name: \hrulefill \\[\namedata@vspace]
%              Student Number: \hrulefill \\[\namedata@vspace]
%              TA: \hrulefill \\[\namedata@vspace]
%              Date: \hrulefill}\end{verbatim}
% Where "\namedata@vspace" is the length set by the switch
% "\StudentInfoLineSpacing".
%
% For those who don't know, "\@@line" is the \LaTeXe\ equivalent of Plain \TeX's
% "\line" command.  If you wish to use "\@@line" in defining your own \textsf{examtop}
% or \textsf{keytop}, please be sure to use "\makeatletter" and "\makeatother"
% in the appropriate places, otherwise you could get errors.
%
% \DescribeEnv{keytop}
% Any material enclosed in the \textsf{keytop} environment will be
% typeset on the next page following the \textsf{frontmatter} text, if a key
% is being created, and will be omitted if an exam is being created.  The
% default settings of the \textsf{examdesign} class use this environment to create
% the top of an answer key as follows:
%  \begin{verbatim}
%  \begin{keytop}
%  \@@line{\hfill \Huge Answer Key
%          for Exam \fbox{\textsf{\Alph{version}}}\fi \hfill}
%          \bigskip
%  \end{keytop}\end{verbatim}
%
% \DescribeEnv{exampreface}
% Any material enclosed in the \textsf{exampreface} environment will be
% typeset on the same page as the \textsf{examtop} text, right beneath it but
% before any sections of the exam are included.  In case you are wondering what
% this environment does that cannot be done with \textsf{examtop}, your question
% is very good.  The difference is this: The material enclosed in the
% \textsf{exampreface} environment is sensitive to whether the \texttt{twocolumn}
% option was given.  If the \texttt{twocolumn} option \emph{was} given, then
% the material in \textsf{exampreface} will be typeset at the start of the
% first column. By contrast, the material given in \textsf{examtop} is
% \emph{always} typeset in a single column spanning the entire textwidth of
% the page, at the top of the exam. 
%
% \DescribeEnv{keypreface}
% Any material enclosed in the \textsf{keypreface} environment will be
% typeset on the same page as the \textsf{keytop} text, right beneath it but
% before any sections are included.  As described above, if the
% \texttt{twocolumn} option \emph{was} given, then the material
% in \textsf{keypreface} will be typeset at the start of the first column. By
% contrast, the material given in \textsf{keytop} is \emph{always} typeset in a single
% column spanning the entire textwidth of the page, at the top of the key.
%
% \DescribeEnv{examclosing}
% Essentially the same as the \textsf{exampreface} environment, with the
% obvious change that the enclosed material is typeset at the end of the exam,
% on the same page as the rest of the exam (if possible), but before the
% \textsf{endmatter}.  (Any material enclosed in an \textsf{endmatter}
% environment is typeset on a page by itself after the exam or key.)
%
% \DescribeEnv{keyclosing}
% Essentially the same as the \textsf{keypreface} environment, with the
% obvious change that the enclosed material is typeset at the end of the key,
% on the same page as the rest of the key (if possible), but before the
% \textsf{endmatter}.
%
%
% \subsection{Special Formatting Macros}
%
% Due to the special way the \textsf{examdesign} class constructs a document (see the
% technical notes in section \ref{TN}), several special macros are provided
% which enable the user to easily take advantage of the way exams are
% constructed.  In addition, there are several lengths which the user can modify
% to further customize the appearance of the exam.
%
% \DescribeMacro{\exam}
% The "\exam" macro takes one argument, which can be any command or
% sequence of commands subject to the constraints described in the technical
% notes.  The argument will be included wherever it is if and only if an exam is
% being made, otherwise the argument will be ignored.
%
% \DescribeMacro{\key}
% The "\key" macro takes one argument, which can be any command or
% sequence of commands subject to the constraints described in the technical
% notes.  The argument will be included wherever it is if and only if a key is
% being made, otherwise the argument will be ignored.
%
% \DescribeMacro{\examvspace}
% Just like the ordinary \LaTeX\ command "\vspace", with the exception that the vertical
% space is included if and only if an exam is being typeset.
%
% \DescribeMacro{\examvspace*}
% Just like the ordinary \LaTeX\ command "\vspace*", with the exception that the vertical
% space is included if and only if an exam is being typeset.
%
% \DescribeMacro{\examhspace}
% Just like the ordinary \LaTeX\ command "\hspace", with the exception that the
% horizontal space is included if and only if an exam is being typeset.
%
% \DescribeMacro{\examhspace*}
% Just like the ordinary \LaTeX\ command "\hspace*", with the exception that the
% horizontal space is included if and only if an exam is being typeset.
%
% \DescribeMacro{\keyvspace}
% Just like the ordinary \LaTeX\ command "\hspace*", with the exception that the
% horizontal space is included if and only if a key is being typeset.
%
% \DescribeMacro{\keyvspace*}
% Just like the ordinary \LaTeX\ command "\hspace*", with the exception that the
% horizontal space is included if and only if a key is being typeset.
%
% \DescribeMacro{\keyhspace}
% Just like the ordinary \LaTeX\ command "\hspace*", with the exception that the
% horizontal space is included if and only if a key is being typeset.
%
% \DescribeMacro{\keyhspace*}
% Just like the ordinary \LaTeX\ command "\hspace*", with the exception that the
% horizontal space is included if and only if a key is being typeset.
%
% \DescribeMacro{\word}
% The last new command (in version 1.02) is "\word".  This command
% allows the user to insert inessential changes in the wording of a question
% between various versions of the exam.  Thus, not only will the order of the
% questions be different, but the actual \emph{wording} of the questions will be
% slightly different as well.  For example, if one includes:
% \begin{verbatim}
% \begin{question}
% Is the most well-known \word{{astronomer} {wrestler} {physicist}} in
% the world \word{{Carl Sagan} {Rowdy Roddy Piper} {Albert Einstein}}?
% \end{question}\end{verbatim}
% in an exam, the first version of the exam will include the question
% \begin{enumerate}
% \item Is the most well-known astronomer in the world Carl Sagan?
% \end{enumerate}
% but the second version of the exam will include
% \begin{enumerate}
% \item Is the most well-known wrestler in the world Rowdy Roddy Piper?
% \end{enumerate}
% and the third version of the exam will include
% \begin{enumerate}
% \item Is the most well-known physicist in the world Albert Einstein?
% \end{enumerate}
%
% Let's call the argument of "\word" an \emph{option-list}.  If one requests
% more versions of an exam than options in an option-list, "\word" will behave
% as if the option list you gave ``wrapped.''  That is, if one requests five
% copies of an exam but includes "\word{{A} {B}}" in the input file, "A" will be
% printed on the fifth version of the exam, because "\word" will act as if the
% option list was "\word{{A} {B} {A} {B} {A} {B}}", where "A" is the fifth
% option.
%
% If it turns out that your exams are becoming absolutely unreadable through
% excessive use of "\word", chances are you are not introducing inessential
% changes in the wording of a question, but are actually trying to write two
% different exams simultaneously.
%
% \subsection{Length Parameters}
%
% The following table describes lengths that the user may set to whatever value
% they wish, and the effect that it has upon the appearance of the exam (or
% key).
%
% \DeleteShortVerb{\|}
% \begin{center}
% \begin{tabular}{|l|l|p{1.8in}|} \hline
% {\large Length} & {\large Initial Value} & {\large Effect} \rule{0pt}{13pt} \\ \hline
% "\beforesectsep" & "\medskipamount"         & Space inserted before the
%                                               current section heading. \\
% "\aftersectsep"  & "\medskipamount"         & Space inserted after the
%                                               current section heading. \\
% "\beforeinstsep" & "\medskipamount"         & Space inserted before any
%                                               instructions for the current
%                                               section are typeset. \\
% "\afterinstsep"  & "\medskipamount-\topsep" & Space inserted after the
%                                               instructions for the current
%                                               section but before the
%                                               questions. \\ \hline
% \end{tabular}
% \end{center}
% \MakeShortVerb{\|}
%
% Given the above parameters, there are two rules that determine whether space
% is inserted.  (1) If the section heading is empty (as might happen if someone
% used a starred environment form and then did not include any optional
% argument), then \emph{neither} "\beforesectsep" \emph{nor} "\aftersecsep" are
% used to insert vertical space, and (2) if there are no instructions for a
% section, then \emph{neither} "\beforeinstsep" \emph{nor} "\afterinstsep" are
% used to insert vertical space.
%
% \section{Miscellaneous Goodies\label{TN}}
%
% \DescribeMacro{\setrandomseed}
% The "\setrandomseed" macro allows you to specify a random seed at the start
% of an exam.  This allows you to have a ``randomized'' exam that keeps
% questions in the same scrambled order across different runs.  
% If you set the random seed, it becomes possible to have crossreferences in
% the exam.  In this release, the ordinary \LaTeX\ commands "\label" and "\ref" 
% work as expected, even across different forms of the test (with the
% questions in a different order).
%
% Math environments using numbered equations should also work correctly
% in this version. 
%
% \DescribeMacro{\pagebreak}
% Having fixed the order of the questions, you might need to introduce a
% pagebreaks at certain points.  In this release, the "\pagebreak" command
% works as expected, with one caveat.  Putting a "\pagebreak" before a section
% will introduce a pagebreak before that section, and putting a "\pagebreak"
% before a question will introduce a pagebreak before that question---no
% matter where on the page that question may be.  
%
% \DescribeMacro{\BreakPageOnVersion}
% If you want to insert pagebreaks before questions or sections conditionally
% on the exam, you will need to use the "\BreakPageOnVersion{}" macro.  This
% macro, which takes a number as its argument, will conditionally insert a
% "\pagebreak" depending on whether number of the version currently being
% typeset equals the version given as an argument to the macro.
%
% \DescribeMacro{\IncludeFromFile}
% \DescribeMacro{\InsertChunk}
% \DescribeEnv{chunk}
% If you need to use any environment which changes the catcodes of the input
% characters (such as the \textsf{verbatim} environment, or XY-pic), you will
% have to employ the following hack.  The reason for this has to do with the
% way \textsf{examdesign} scrambles the order of the questions.  (Basically, each
% question is saved to a macro for later recall.  This assigns each character
% a catcode when it is saved, so it is not possible to change the catcode at a
% later time.)
%
% The hack is the following: create another file called ``foo.tex'' (you can
% name it whatever you like) and put "\IncludeFromFile{foo.tex}" in
% the preamble of your document.  For each bit of text that you need contained
% category code changes (like the \textsf{verbatim} environment) put, in
% foo.tex, a named chunk containing that code.  I.e.,
% \begin{quote}
% "\begin{chunk}{chunk name}"\\
% "\begin{verbatim}"\\
% "  This is some verbatim text"\\
% "  This is more verbatim text"\\
% "\end{verbatim}"\\
% "\end{chunk}"
% \end{quote}
% At the appropriate place in your exam, place "\InsertChunk{chunk name}".
% This macro inserts the named chunk into the exam (or key) at that place.
% All catcode changing commands, environments, etc., should be capable of
% being inserted this way.  (If you find one that doesn't, let me know.)
%
% All of the chunks for a given exam can be stored in the same file---you just
% need to give each chunk a unique name. Only the
% chunk you request will be inserted at a particular point.  If you have a lot
% of chunks, this can slow down considerably the overall processing of the exam,
% but there's no other general solution which will work.
% 
% The chunk names can be descriptive, but don't try anything too fancy.  ``Too
% fancy'' meaning: if it causes an error, it's too fancy. 
%
% \subsection{Bugs}
%
% No known bugs.
%
% \subsection{Changes}
%
% The changes between this version and previous versions are indicated in the
% above sections. Early releases of the \textsf{examdesign} class required that
% instructions 
% for a given section be specified as such by using the "\instructions"
% command. The need for using "\instructions" has been lifted as of release
% 1.02.  The rule is this: all material between the start of a section and
% the first "\begin{question}" or "\begin{block}" will be taken to be the
% instructions for that section.  Old exams created with "\instructions" should
% still work with the new release (the "\instructions" macro has simply been
% redefined as "\relax").
%
% Thus, if you are using
% \textsf{fixed} to format a series of essay questions, you might want to tell
% students that they only need to write answers to two of the following three
% questions.  The instruction text is always
% typeset at the beginning of a section, before any questions are typeset.
%
% \endinput
%
% \section{The Code}
%
% The following code is a bit messy, and the comments don't necessarily match
% up with the code that follows.  I thought it was more important to get the
% new version out than worry about making the source pretty.  If you need to
% figure out how something works, trust the code, not the description!
%
% First just some declarations and stuff that says we are building an
% extension to the ordinary \LaTeX\ class |article|.
%    \begin{macrocode}
%<*header>
\ProvidesFile{examdesign.dtx} 
\NeedsTeXFormat{LaTeX2e} 
\ProvidesClass{examdesign}
\newcounter{passes}\setcounter{passes}{2} 
\newif\ifmcbold \mcboldfalse
\newif\ifmcunder \mcunderfalse 
\newif\ifulem \ulemfalse 
\newif\ifbox \boxtrue
\newif\iftwocolumn \twocolumnfalse 
\gdef\@OneKeyHook{\relax}
%\DeclareOption{mcbold}{\mcboldtrue\boxfalse}
%\DeclareOption{mcunderline}{\mcundertrue \boxfalse}
%\DeclareOption{nokey}{\setcounter{passes}{1}}
%\DeclareOption{onekey}{\setcounter{passes}{2}%
%  \gdef\@OneKeyHook{\setcounter{passes}{1} 
%  \gdef\@OneKeyHook{\relax}}}
\DeclareOption{twocolumn}{\twocolumntrue}
\DeclareOption*{\PassOptionsToClass{\CurrentOption}{article}} 
\ProcessOptions
\LoadClass{article} 

\def\NoKey{\setcounter{passes}{1}}

\def\OneKey{\setcounter{passes}{2}%
  \gdef\@OneKeyHook{\setcounter{passes}{1}
    \gdef\@OneKeyHook{\relax}}}

\def\BoldfaceCorrectMultipleChoiceAnswer{%
  \mcboldtrue
  \boxfalse
  \exam@SetCorrectChoiceFont}

\def\UnderlineCorrectMultipleChoiceAnswer{%
  \mcundertrue
  \boxfalse
  \exam@SetCorrectChoiceFont}

\def\exam@SetCorrectChoiceFont{%
  \ifmcbold 
    \ifmcunder 
    \def\MCfont##1{\textbf{\uline{##1}}}%
    \else 
    \def\MCfont##1{\textbf{##1}}%
    \fi 
  \else 
    \ifmcunder
    \def\MCfont##1{\uline{##1}}%
    \else 
    \def\MCfont##1{\relax{##1}}%
    \fi 
  \fi}

% \ifmcbold 
%   \ifmcunder 
%   \def\MCfont#1{\textbf{\uline{#1}}}
%   \else 
%   \def\MCfont#1{\textbf{#1}} 
%   \fi 
% \else 
%   \ifmcunder
%   \def\MCfont#1{\uline{#1}} 
%   \else 
%   \def\MCfont#1{\relax{#1}} 
%   \fi 
% \fi


\RequirePackage{enumerate} 
\RequirePackage{multicol} 
\RequirePackage{keyval}
%</header>
%    \end{macrocode}
%
% \subsection{The random number generator}
%
% RANDOM.TEX       v.1   (Donald Arseneau)
% Generating "random" numbers in TeX.
%
% \DescribeMacro{\nextrandom}
% Random integers are generated in the range 1 to 2147483646 by the
% macro |\nextrandom|.  The result is returned in the counter |\randomi|.
% Do not change |\randomi| except, perhaps, to initialize it at some
% random value.  If you do not initialize it, it will be initialized
% using the time and date.  (This is a sparse initialization, giving
% fewer than a million different starting values, but you should use
% other sources of numbers if they are available--just remember that
% most of the numbers available to \TeX are not at all random.)
%
% \DescribeMacro{\setrannum}
% The |\nextrandom| command is not very useful by itself, unless you
% have exactly 2147483646 things to choose from.  Much more useful
% is the |\setrannum| command which sets a given counter to a random
% value within a specified range.  There are three parameters:
% |\setrannum{|\meta{counter}|}{|\meta{minimum}|}{|\meta{maximum}|}|.
% For example, to simulate a die-roll: |\setrannum{\die}{1}{6} \ifcase\die|\ldots
%
% \DescribeMacro{\setrandimen}
% \DescribeMacro{\pointless}
% If you need random numbers that are not integers, you will have to
% use dimen registers and |\setrandimen|.  For example, to set a random
% page width: |\setrandimen \hsize{3in}{6.5in}|.  The ``|\pointless|'' macro
% will remove the ``pt'' that TeX gives so you can use the dimensions
% as pure `real' numbers.  In that case, specify the range in pt units.
% For example, |\setrandimen\answer{2.71828pt}{3.14159pt}|
% The answer is |\pointless\answer|.
%
% The random number generator is the one by Lewis, Goodman, and Miller
% (1969) and used as "ran0" in "Numerical Recipies" using Schrage's
% method for avoiding overflows.  The multiplier is $16807 (7^5)$ the
% added constant is 0, and the modulus is $147483647 (2^{31}-1)$  The
% range of integers generated is $1 - 2147483646$.  A smaller range would
% reduce the complexity of the macros a bit, but not much--most of the
% code deals with initialization and type-conversion.  On the other hand,
% the large range may be wasted due to the sparse seed initialization.
%
% I do not include the code in the documentation.
%\iffalse
%    \begin{macrocode}
%<*random>
\newcount\randomi % the random number seed (while executing)
\global\randomi\catcode`\@  % scratch variable during definitions
\catcode`\@=11
% The following macro isn't part of Arseneau's original TeX package,
% but it comes in handy when users want to recover a particular run.
\def\setrandomseed#1{\global\randomi#1\relax
  \message{Randomizer initialized to \the\randomi.}%
  \nextrandom \nextrandom \nextrandom}

\def\nextrandom{\begingroup
 \ifnum\randomi<\@ne % then initialize with time
    \global\randomi\time
    \global\multiply\randomi388 \global\advance\randomi\year
    \global\multiply\randomi31 \global\advance\randomi\day
    \global\multiply\randomi97 \global\advance\randomi\month
    \message{Randomizer initialized to \the\randomi.}%
    \nextrandom \nextrandom \nextrandom
 \fi
 \count@ii\randomi
 \divide\count@ii 127773 % modulus = multiplier * 127773 + 2836
 \count@\count@ii
 \multiply\count@ii 127773
 \global\advance\randomi-\count@ii % random mod 127773
 \global\multiply\randomi 16807
 \multiply\count@ 2836
 \global\advance\randomi-\count@
 \ifnum\randomi<\z@ \global\advance\randomi 2147483647\relax\fi
 \endgroup
}

\countdef\count@ii=2 % use only in boxes!
\ifx\@tempcnta\undefined \csname newcount\endcsname \@tempcnta \fi
\ifx\@tempcntb\undefined \csname newcount\endcsname \@tempcntb \fi

\def\setrannum#1#2#3{% count register, minimum, maximum
 \@tempcnta#3\advance\@tempcnta-#2\advance\@tempcnta\@ne
 \@tempcntb 2147483645 %  =  m - 2  =  2^{31} - 3
 \divide\@tempcntb\@tempcnta
 \getr@nval
 \advance\ranval#2\relax
 #1\ranval
}

\def\setrandim#1#2#3{% dimen register, minimum length, maximum length
 \dimen@#2\dimen@ii#3\relax
 \setrannum\ranval\dimen@\dimen@ii
 #1\ranval sp\relax
}

\def\getr@nval{% The values in \@tempcnta and \@tempcntb are parameters
 \nextrandom
 \ranval\randomi \advance\ranval\m@ne \divide\ranval\@tempcntb
 \ifnum\ranval<\@tempcnta\else \expandafter\getr@nval \fi
}

\def\pointless{\expandafter\PoinTless\the}
{\catcode`p=12 \catcode`t=12
\gdef\PoinTless#1pt{#1}}

\catcode`\@=\randomi
\global\randomi=0
\newcount\ranval
%</random>
%    \end{macrocode}
%\fi
%
% \subsection{Incorporating the package \textsf{ulem}}
% This code is not listed in the documentation.
%\iffalse
%    \begin{macrocode}
%<*ulem>
\expandafter \ifx \csname UL@box\endcsname \relax \else
 \typeout{ulem.sty refuses to load twice. }\endinput \fi
\chardef\ULthickness\catcode\string`\@ % hold catcode temporarily
\catcode\string`\@=11
\def\uline{\relax \ifmmode\expandafter\underline
           \else \bgroup\expandafter\ULset\fi}
\newbox\UL@box
\newbox\UL@hyphenbox
\newskip\UL@skip
\newcount\UL@pe
\let\LA@space\ \let\LA@hskip\hskip
\def\UL@end *{\relax\relax}% something harmless unlikely to be found elsewhere
\def\ULset{\ifdim\ULdepth=\maxdimen\setbox\UL@box\hbox{(j}\ULdepth\dp\UL@box\fi
  \def\UL@leadtype{\leaders \hrule \@height\dimen@ \@depth\ULdepth}%
  \ifmmode \ULdepth-4\p@ \fi
  \dimen@-\ULdepth \advance\dimen@\ULthickness \ULon}
\def\ULon{\ifmmode \expandafter\UL@onmath\else \expandafter\UL@on\fi}
\def\UL@on#1{\leavevmode\UL@ender \let\UL@on\UL@onin
  \everymath{\UL@hrest}\everyvbox{\UL@hrest}\let\hskip\UL@hskip
  \let\\\UL@cr \let\-\UL@dischyp \let\newline\UL@newline \let\ \UL@space
  \def\hfil{\hskip\z@ plus1fil\relax}\def\hfill{\hskip\z@ plus1fill\relax}%
  \def\hss{\hskip\z@ plus1filminus1fil\relax}\let\penalty\UL@penalty
  \UL@word\@empty#1\xdef\UL@spfactor{\the\spacefactor} \UL@end * }
\def\UL@onin#1{\leavevmode\UL@ender % when nested, do multiple underlining
  \dimen@\ULthickness \advance\ULdepth\thr@@\dimen@ \advance\dimen@-\ULdepth
  \setbox\UL@box\hbox{{#1}}\let\UL@start\relax\UL@putbox\egroup}
\def\UL@onmath#1{\UL@ender\mathord{\UL@hrest\mathop{\kern\z@#1}\limits\sb
  {\UL@leadtype\LA@hskip\p@ plus1fill}}\egroup}
\def\UL@unegroup{}
\gdef\UL@ender{}
\def\UL@swender{\ifnum`{=\z@\fi\aftergroup}\gdef\UL@ender{}}
\def\UL@word#1 {\expandafter\UL@start#1 %
  \expandafter\ifx\expandafter\UL@end#1\egroup\egroup
   \unskip \unskip \unskip % remove extra leader at end
   \spacefactor\UL@spfactor \let\UL@word\egroup
  \else % not finished
   \ifmmode\else \ifdim\lastskip=\z@\else % allow syntactical spaces
     \global\UL@skip\lastskip \unskip
     \UL@stop \UL@leaders
   \fi\fi
  \fi \UL@word\@empty}% \@empty preserves braces in param
\def\UL@start{\setbox\UL@box\hbox\bgroup\everyhbox{\UL@hrest}%
  \let\UL@start\@empty \def\UL@unegroup{\bgroup\bgroup}\let\UL@leadtype\@empty
  \bgroup \kern-3sp\kern3sp % kerns so I can test for beginning of list
  \if@ignore \global\@ignorefalse \ignorespaces \fi}
\def\UL@stop{\global\UL@pe\lastpenalty \unpenalty % penalty in \UL@pe
  \ifnum\lastkern=\thr@@ \egroup\egroup % Nothing in hbox...but make sure:
    \ifdim\wd\UL@box=\z@ \else \UL@putbox \fi % something in box so print it
  \else  \egroup\egroup \UL@putbox % something in box so print it
  \fi \ifnum\UL@pe=\z@ \else \LA@penalty\UL@pe \fi % use penalty from inside box
  \UL@unegroup}
\def\UL@putbox{\ifx\UL@start\@empty \else % not inner
  \vrule\@width\z@ \LA@penalty\@M
  {\UL@skip\wd\UL@box \UL@leaders \kern-\UL@skip}%
  \box\UL@box  \fi}
\newdimen\UL@pixel \UL@pixel=1in \divide\UL@pixel 300
\def\UL@leaders{{\LA@hskip-\UL@pixel \advance\UL@skip\tw@\UL@pixel
   \UL@leadtype\LA@hskip\UL@skip  \LA@hskip-\UL@pixel}}
\def\UL@hrest{\let\ \LA@space \let\-\@empty \let\penalty\LA@penalty}
\def\UL@space{\LA@space \global\UL@skip\lastskip \unskip \UL@reskip}% \
\def\UL@dischyp{\global\setbox\UL@hyphenbox\hbox
   {\ifnum \hyphenchar\font<\z@ \string-\else \char\hyphenchar\font \fi}%
   \kern\wd\UL@hyphenbox \LA@penalty\@M \UL@stop \kern-\wd\UL@hyphenbox
   \discretionary{\box\UL@hyphenbox}{}{}\UL@start}
\let\LA@penalty\penalty
\def\UL@penalty{\relax\ifhmode \afterassignment\UL@@penalty\count@
  \else\LA@penalty\fi}
\def\UL@@penalty{\LA@penalty \ifnum\count@=\z@
  \@ne \else \count@ \fi % zero penalty => no penalty, so use 1 instead.
  \UL@stop \UL@start}
\def\UL@hskip{\ifx\ \LA@space \LA@hskip \else
  \afterassignment\UL@reskip \global\UL@skip \fi}
\def\UL@reskip{\UL@stop \UL@leaders \UL@start}
\def\UL@cr{\unskip \ifx\ \LA@space \let\UL@vad\@gobble
  \else \UL@stop \unskip\unskip\unskip \let\UL@vad\vadjust \fi
  \@ifstar{\UL@vad{\LA@penalty\@M}\UL@cra}\UL@cra}
\def\UL@cra{\@ifnextchar[\UL@crb\UL@newline}
\def\UL@crb[#1]{\UL@vad{\vskip#1}\UL@newline}
\def\UL@newline{\ifx\UL@start\@empty % (\UL@cr may have \UL@stop-ed already)
  \unskip \ifx\ \LA@space \else \UL@stop \unskip\unskip\unskip \fi\fi
  \LA@hskip \z@\@plus.0001fil\LA@penalty -\@M \UL@start}
\newbox\ULC@box
\def\markoverwith#1{\setbox\ULC@box\hbox{{#1}}\UL@pixel.5\wd\ULC@box
   \ifmmode \setbox\ULC@box\hbox{\raise1.4ex\box\ULC@box}%
    \dp\ULC@box-1.4ex\ht\ULC@box\z@ \def\UL@leadtype{\cleaders\copy\ULC@box}%
   \else
    \def\UL@leadtype{\leaders\copy\ULC@box}%
   \fi}
\def\uwave{\bgroup \markoverwith{\lower3.5\p@\hbox{\sixly
  \kern-.21\p@ \char58 \kern-.21\p@}}\ULon}
\font\sixly=lasy6 % does not re-load if already loaded, so no memory problem.
\def\uuline{\bgroup \markoverwith{\hbox
 {\kern-.03em\vtop{\kern.2ex\hrule width.2em\kern1.1\p@
 \hrule}\kern-.03em}}\ULon}
\def\sout{\bgroup \ULdepth=-.5ex \ULset}
\def\xout{\bgroup \markoverwith{\kern-.1em{/}\kern-.1em}\ULon}
\def\useunder#1#2#3{\relax
  \ifx\relax#2\relax\else % declaration command given
    \def#2{\def\@tempa{#1}\global\let\UL@ender\UL@swender
    \expandafter\@tempa\expandafter{\ifnum\z@=\string`}\fi}%
    \MakeRobust{#2}\fi
  \ifx\relax#3\relax\else % argumentative command
    \def#3{#1}\MakeRobust{#3}\fi}
\expandafter\ifx \csname @ifundefined\endcsname \relax
  \def\@height{height}
  \def\@depth{depth}
  \def\@width{width}
  \def\@empty{}
  \def\@gobble#1{}
  \def\MakeRobust#1{}
  \csname newif\expandafter\endcsname \csname if@ignore\endcsname
\else
 \def\MakeRobust#1{\expandafter\let
   \csname \expandafter\@gobble\string#1 \endcsname= #1%
  \edef#1{\noexpand\protect \expandafter\noexpand
   \csname\expandafter\@gobble\string#1 \endcsname}
  }
  \MakeRobust\uline
  \MakeRobust\uuline
  \MakeRobust\uwave
  \MakeRobust\sout
  \MakeRobust\xout
  \let\LA@em\em   \let\LA@emph\emph
  \expandafter\let\expandafter\LA@Pem \csname em \endcsname
  \expandafter\let\expandafter\LA@Pemph \csname emph \endcsname
  \def\ULforem{\useunder{\uline}{\em}{\emph}}
  \def\normalem{\let\em\LA@em \let\emph\LA@emph
    \expandafter\let\csname em \endcsname\LA@Pem
    \expandafter\let\csname emph \endcsname\LA@Pemph}
  \ULforem  %  default is to use underlining for \em,
\fi
\catcode`@=\ULthickness % Scratch meaning: restore catcode of @
\def\ULthickness{.4pt}% can change this with \renewcommand
\newdimen\ULdepth  \ULdepth=\maxdimen

\normalem
%</ulem>
%    \end{macrocode}
%\fi
%
% \subsection{The code for the \textsf{examdesign} class}
%
% We now define an environment \textsf{question} that saves the surrounded text
% to a macro |\sec@|\meta{section-number}|@qst|\meta{question-number}, where
% \meta{section-number} and \meta{question-number} are expressed as lower case
% roman numerals (this way they can be part of the macro name).  Note that we
% must use lower case roman numerals since, I believe, \LaTeXe\ the |\Roman|
% command uses the |\uppercase| command, which is not expanded fully until
% further along in \TeX's gullet than we want.
%
% First, define some new counters and token lists.
%    \begin{macrocode}
%<*othercode>
\newcounter{questionnumber}
\setcounter{questionnumber}{0}
\setcounter{section}{0}
\newcounter{extract}
\newcounter{question} \setcounter{question}{1}
\newcounter{length} \setcounter{length}{0}
\newcounter{temp} \setcounter{temp}{1}
\newcounter{wordtemp} \setcounter{wordtemp}{1}
\newcounter{form}
\newcounter{version}
\newcounter{current@section}
\newcounter{current@question}
\newcounter{innerloop}
\newcounter{sect}
\newcounter{sectionindex}
\newcounter{answer}
\newcounter{choice}
 \setcounter{choice}{1}
\newcounter{mcquestion} \setcounter{mcquestion}{1}
\newcounter{mcchoice} \setcounter{mcchoice}{1}
\newcounter{savelistcnt}
\newlength{\blank@length}
\newlength{\temp@length}
\newlength{\temp@@length}
\newlength{\namedata@vspace}
\newlength{\beforesectsep}
 \setlength{\beforesectsep}{\medskipamount}
\newlength{\aftersectsep}
 \setlength{\aftersectsep}{\medskipamount}
\newlength{\beforeinstsep}
 \setlength{\beforeinstsep}{\medskipamount}
\newlength{\afterinstsep}
 \setlength{\afterinstsep}{\medskipamount}
 \addtolength{\afterinstsep}{-\topsep}
\newif\ifchoice \choicefalse
\newif\ifanswer
\newif\ifnotfixed
\newif\ifcblanks \cblanksfalse
\newif\ifpblanks \pblankstrue
\newif\ifidentifier \identifierfalse
\newif\ifFirstQuestionInBlock \FirstQuestionInBlocktrue
\newif\ifcontinuousnumbering \continuousnumberingtrue
\newif\ifresetcounter\resetcounterfalse
\global\newtoks\temptoks
\global\newtoks\temp@toks
%    \end{macrocode}
% |\setrannum| requires a ``traditional'' counter (i.e., a non-\LaTeX counter (i.e., a
% counter whose name doesn't begin with |c@|.
%    \begin{macrocode}
\newcount\exam@c
\newcount\exam@n
\newcount\exam@m
\setrannum{\exam@m}{1}{10}
\newtoks\@toklist
%    \end{macrocode}
%
% \begin{environment}{section}
% Questions naturally occur in blocks which have, more-or-less, the same
% general format.  For example, one will naturally have a block of true/false
% questions together, followed by a block of short answer, and then a
% block of essay questions.  We don't want to mix questions between blocks (so
% that the true/false questions are interspersed between the short answer
% questions).  Thus we need to index the questions by section type as well as
% question number.
%
%    \begin{macrocode}
\def\section{\stepcounter{section}
  \setcounter{question}{1}}
%    \end{macrocode}
%
% \begin{macro}{\get@instructions}
% This macro collects everything between it and the next |\begin| token.  It then
% calls the |\find@instructions@end| macro to check whether the |\begin| token
% belongs to the \textsf{question} environment or some other environment.
%    \begin{macrocode}
\long\def\get@instructions#1\begin{\@toklist=\expandafter{\the\@toklist#1}%
  \find@instructions@end}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\find@instructions@end}
% If we are at |\begin{question}| or |\begin{block}|
% save everything to the appropriately named
% macro.  Otherwise, continue collecting tokens
%    \begin{macrocode}
\def\find@instructions@end#1{%
  \def\@tempa{#1}%
  \ifx\@tempa\@Q
    \let\@next=\Qwrap@up@instructions
  \else
    \ifx\@tempa\@Block
      \let\@next=\Bwrap@up@instructions
    \else
      \@toklist=\expandafter{\the\@toklist\begin{#1}}%
      \let\@next=\get@instructions
   \fi\fi\@next}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\Qwrap@up@instructions}
% Use the counters |section| and |question| to save the collected token list
% in a macro named as described earlier.
%    \begin{macrocode}
\def\Qwrap@up@instructions{%
 \expandafter\xdef\csname inst@sec\thesection\endcsname{\the\@toklist}
 \begin{question}}
% \end{macro}
%
% \begin{macro}{\Bwrap@up@instructions}
% Use the counters |section| and |question| to save the collected token list
% in a macro named as described earlier.
%    \begin{macrocode}
\def\Bwrap@up@instructions{%
 \expandafter\xdef\csname inst@sec\thesection\endcsname{\the\@toklist}
 \begin{block}}
% \end{macro}
%
% \begin{macro}{\endsection}
% At the end of a section, make a list of question indexes.
%    \begin{macrocode}
\def\endsection{\make@qlist}
%    \end{macrocode}
% \end{macro}
% \end{environment}
%
% \begin{macro}{\make@qlist}
% As described below, we need to create a token list containing the index
% numbers of each question.  This is done by a while-loop as follows:
%    \begin{macrocode}
\def\make@qlist{%
  \setcounter{temp}{1}
  \edef\@tempa{\thesection}%
%    \end{macrocode}
% We need to create a new token list to hold the question indexes.  While
% we are at it, might as well create the token list to hold the rearranged
% question indexes.  This saves us having to include a while-loop later.
%    \begin{macrocode}
  \global\expandafter\newtoks\csname qlist@sec\@tempa\endcsname
  \global\expandafter\newtoks\csname qlist@scr@sec\@tempa\endcsname
  \global\csname qlist@scr@sec\@tempa\endcsname={}
%    \end{macrocode}
% Now being a while-loop where each pass through the loop appends
% an index number to the question list.
%    \begin{macrocode}
  \@whilenum \value{question}>1 \do{%
    \edef\@questionnumber{\arabic{temp}}%
    \expandafter\append\expandafter{\@questionnumber}%
       {\csname qlist@sec\@tempa\endcsname}
    \addtocounter{question}{-1}
    \addtocounter{temp}{1}}}
%    \end{macrocode}
% \end{macro}
%
% \subsection{The \textsf{question} environment}
%
% \begin{macro}{\question}
% We define |\question| so that it reads in and saves the surround text
% to a macro like in the \textsf{tabularx} environment.
%    \begin{macrocode}
\def\question{%
  \@toklist={\exam@ExecuteIfDefined{sec\thesection q\thequestion break}}%
%   \@ifundefined{sec\thesection q\thequestion break}
%     {\@toklist={}}
%     {\expandafter\@toklist\expandafter{\csname 
%         sec\thesection q\thequestion break\endcsname}}  
 \get@question@body}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\endquestion}
% Increment the question counter so that the next question will be assigned the
% right name
%    \begin{macrocode}
\def\endquestion{\stepcounter{question}}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\get@question@body}
% This macro collects everything between it and the next |\end| token.  It then
% calls the |\find@question@end| macro to check whether the |\end| token
% belongs to the \textsf{question} environment or some other environment.
%    \begin{macrocode}
\long\def\get@question@body#1\end{\@toklist=\expandafter{\the\@toklist#1}%
  \find@question@end}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\find@question@end}
% If we are at |\end{question}|, save everything to the appropriately named
% macro.  Otherwise, continue collecting tokens
%    \begin{macrocode}
\def\find@question@end#1{%
  \def\@tempa{#1}%
   \ifx\@tempa\@Q\expandafter\wrap@up
    \else\@toklist=\expandafter{\the\@toklist\end{#1}}%
     \expandafter\get@question@body\fi}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\wrap@up}
% Use the counters |section| and |question| to save the collected token list
% in a macro named as described earlier.
%    \begin{macrocode}
\def\exam@ExecuteIfDefined#1{%
  \@ifundefined{#1}%
    {}%
    {\csname #1\endcsname}} 

\def\wrap@up{%
 \edef\@tempa{\thesection}%
 \edef\@tempb{\thequestion}%
 \ifx\@currentsectionname\@truefalse
   \expandafter\xdef\csname sec\@tempa @qst\@tempb\endcsname{%
     \noexpand\exam@ExecuteIfDefined{sec\thesection q\thequestion break}%
     \noexpand\exam@TFItem
     \noexpand\@beforequestionhook
     \noexpand\exam@TFGobbleQuestionHook
     \the\@toklist
     \noexpand\@endquestionhook}
   \else
     \ifx\@currentsectionname\@MC
       \expandafter\xdef\csname sec\@tempa @qst\@tempb\endcsname{%
         \noexpand\exam@ExecuteIfDefined{sec\thesection q\thequestion break}%
         \noexpand\@mcitem
         \noexpand\@beforequestionhook
         \noexpand\exam@MCGobbleQuestionHook
         \the\@toklist
         \noexpand\exam@EndChoiceListIfNecessary
         \noexpand\@mcenditem
         \noexpand\@endquestionhook}
       \else
         \ifx\@currentsectionname\@shortanswer
           \expandafter\xdef\csname sec\@tempa @qst\@tempb\endcsname{%
             \noexpand\exam@ExecuteIfDefined{sec\thesection q\thequestion break}%
             \noexpand\item
             \noexpand\@beforequestionhook
             \noexpand\exam@SAGobbleQuestionHook
             \the\@toklist
             \noexpand\@endquestionhook}
         \else
           \ifx\@currentsectionname\@fillin
             \expandafter\xdef\csname sec\@tempa @qst\@tempb\endcsname{%
               \noexpand\exam@ExecuteIfDefined{sec\thesection q\thequestion break}%
               \noexpand\item
               \noexpand\@beforequestionhook
               \noexpand\exam@FillinGobbleQuestionHook
               \the\@toklist
               \noexpand\@endquestionhook}
           \else
             \expandafter\xdef\csname sec\@tempa @qst\@tempb\endcsname{%
               \noexpand\exam@ExecuteIfDefined{sec\thesection q\thequestion break}%
               \noexpand\item
               \noexpand\@beforequestionhook
               \the\@toklist
               \noexpand\@endquestionhook}
   \fi\fi\fi\fi
   \end{question}}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\@Q}
% Store the name of the environment for comparison in |\find@question@end|
%    \begin{macrocode}
\def\@Q{question}
%    \end{macrocode}
% \end{macro}
%
% \subsection{The \textsf{block} environment}
%
% \begin{macro}{\block}
% This macro provides a ``block'' environment that prevents the questions
% it surrounds from being rearranged.
%
% The |\global\@newlistfalse| is required to prevent \LaTeX\ from barfing
% if the random rearrangement places the block at the beginning of the
% question list, in which case the |\end{enumerate}| would close a
% |\begin{enumerate}| without an intervening |\item|.
%    \begin{macrocode}
\newcounter{first}
\newcounter{last}
\def\exam@GobbleBlockInstructionsHook{\relax}
\newcounter{exam@tf}\setcounter{exam@tf}{0}

\newcommand{\block}[1][]{%
  % Break only if a pagebreak preceeded the block
  \@toklist={\exam@ExecuteIfDefined{sec\current@section q\current@question break}}%
  \@toklist=\expandafter{\the\@toklist
    \global\@newlistfalse
    \setcounter{questionnumber}{\value{enumi}}}%
  \ifx\@currentsectionname\@truefalse
    \@toklist=\expandafter{\the\@toklist
      \end{list}%
      \setcounter{exam@tf}{0}%
      \exam@TrueFalseGobbleBlockInstructionsHook}%
  \else
    \@toklist=\expandafter{\the\@toklist
      \end{enumerate}\exam@GobbleBlockInstructionsHook}%
  \fi
  \setkeys{block}{#1}%
  \exam@GetBlockToFirstQuestion}

\define@key{block}{questions}[0]{%
  \@toklist=\expandafter{\the\@toklist
    \setcounter{first}{\value{enumi}}
    \stepcounter{first}
    \setcounter{last}{\value{enumi}}
    \addtocounter{last}{#1}}}

\long\def\exam@GetBlockToFirstQuestion#1\begin{%
  \@toklist=\expandafter{\the\@toklist #1}%
  \exam@CheckForQuestion}

\def\exam@CheckForQuestion#1{%
  \def\@tempa{#1}%
  \ifx\@tempa\@Q
    \ifx\@currentsectionname\@truefalse
      \@toklist=\expandafter{\the\@toklist
        \exam@StartTrueFalseList
        \begin{question}}
    \else
      \@toklist=\expandafter{\the\@toklist
        \begin{enumerate}
        \setcounter{enumi}{\value{questionnumber}}%
        \begin{question}}
    \fi
    \expandafter\get@block@body
  \else
    \expandafter\exam@GetBlockToFirstQuestion
  \fi}

\long\def\get@block@body#1\end{\@toklist=\expandafter{\the\@toklist #1}%
  \find@block@end}
\def\find@block@end#1{%
  \def\@tempa{#1}%
   \ifx\@tempa\@Block\expandafter\block@wrap@up
    \else\@toklist=\expandafter{\the\@toklist\end{#1}}%
     \expandafter\get@block@body\fi}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\block@wrap@up}
%    \begin{macrocode}
\def\block@wrap@up{%
 \edef\@tempa{\thesection}%
 \edef\@tempb{\thequestion}%
 \expandafter\xdef\csname sec\@tempa @qst\@tempb\endcsname{%
   \noexpand\@startofblockhook
   \the\@toklist
   \noexpand\@endofblockhook}%
 \end{block}}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\@startofblockhook}
%    \begin{macrocode}
\def\@startofblockhook{\relax}
%    \end{macrocode}
% \end{macro}
%
%
% \begin{macro}{\@endofblockhook}
%    \begin{macrocode}
\def\@endofblockhook{\relax}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\@Block}
%    \begin{macrocode}
\def\@Block{block}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\endblock}
%    \begin{macrocode}
\def\endblock{\stepcounter{question}}
%    \end{macrocode}
% \end{macro}
%
% \subsection{The \textsf{frontmatter} environment}
%
% \begin{macro}{\frontmatter}
% We define "\frontmatter" so that it reads in and saves the surround text
% to a macro like in the \textsf{tabularx} environment.
%    \begin{macrocode}
\def\frontmatter{\@toklist={}
 \get@frontmatter@body}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\endfrontmatter}
% This doesn't need to do anything.
%    \begin{macrocode}
\def\endfrontmatter{\relax}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\get@frontmatter@body}
% This macro collects everything between it and the next |\end| token.  It then
% calls the |\find@frontmatter@end| macro to check whether the |\end| token
% belongs to the \textsf{frontmatter} environment or some other environment.
%    \begin{macrocode}
\long\def\get@frontmatter@body#1\end{%
  \@toklist=\expandafter{\the\@toklist#1}%
  \find@frontmatter@end}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\find@frontmatter@end}
% If we are at |\end{frontmatter}|, save everything to the appropriately named
% macro.  Otherwise, continue collecting tokens
%    \begin{macrocode}
\def\find@frontmatter@end#1{%
  \def\@tempa{#1}%
  \ifx\@tempa\@Frontmatter
    \expandafter\frontmatter@wrap@up
  \else
    \@toklist=\expandafter{\the\@toklist\end{#1}}%
    \expandafter\get@frontmatter@body
  \fi}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\frontmatter@wrap@up}
% Save all the scanned text into an appropriately named macro for later
% recall.
%    \begin{macrocode}
\def\frontmatter@wrap@up{%
  \xdef\@frontmattertext{\the\@toklist}
  \end{frontmatter}}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\@Frontmatter}
% Store the name of the environment for comparison in |\find@frontmatter@end|
%    \begin{macrocode}
\def\@Frontmatter{frontmatter}
%    \end{macrocode}
% \end{macro}
%
% \subsection{The \textsf{endmatter} environment}
%
% The \textsf{endmatter} environment behaves just like the \textsf{frontmatter}
% environment, except the enclosed text is saved in a differently-named macro.
%
% \begin{macro}{\endmatter}
% We define "\endmatter" so that it reads in and saves the surround text
% to a macro like in the \textsf{tabularx} environment.
%    \begin{macrocode}
\def\endmatter{\@toklist={}
 \get@endmatter@body}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\endendmatter}
% This doesn't need to do anything.
%    \begin{macrocode}
\def\endendmatter{\relax}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\get@endmatter@body}
% This macro collects everything between it and the next |\end| token.  It then
% calls the |\find@endmatter@end| macro to check whether the |\end| token
% belongs to the \textsf{endmatter} environment or some other environment.
%    \begin{macrocode}
\long\def\get@endmatter@body#1\end{%
  \@toklist=\expandafter{\the\@toklist#1}%
  \find@endmatter@end}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\find@endmatter@end}
% If we are at |\end{endmatter}|, save everything to the appropriately named
% macro.  Otherwise, continue collecting tokens
%    \begin{macrocode}
\def\find@endmatter@end#1{%
  \def\@tempa{#1}%
  \ifx\@tempa\@Endmatter
    \expandafter\endmatter@wrap@up
  \else
    \@toklist=\expandafter{\the\@toklist\end{#1}}%
    \expandafter\get@endmatter@body
  \fi}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\endmatter@wrap@up}
% Save the scanned text to an appropriately named macro for later recall.
%    \begin{macrocode}
\def\endmatter@wrap@up{%
  \xdef\@endmattertext{\the\@toklist}
  \end{endmatter}}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\@Endmatter}
% Store the name of the environment for comparison in |\find@endmatter@end|
%    \begin{macrocode}
\def\@Endmatter{endmatter}
%    \end{macrocode}
% \end{macro}
%
% We now initialize "\@frontmattertext" and "\@endmattertext" to a randomly
% chosen garbage value that no user will ever use.  This enables us to test in
% the main loop whether the user has used \textsf{frontmatter} or
% \textsf{endmatter} in her document.
%
%    \begin{macrocode}
\begin{frontmatter}
  \relax
\end{frontmatter}
%    \end{macrocode}
%
%    \begin{macrocode}
\begin{endmatter}
  \relax
\end{endmatter}
%    \end{macrocode}
%
% \subsection{The \textsf{exampreface} environment}
%
% The \textsf{exampreface} environment behaves just like the \textsf{frontmatter}
% environment, except the enclosed text is saved in a differently-named
% macro---one that will be printed just beneath any text specified in the
% \textsf{examtop} environment.
%
% \begin{macro}{\exampreface}
% We define "\exampreface" so that it reads in and saves the surround text
% to a macro like in the \textsf{tabularx} environment.
%    \begin{macrocode}
\def\exampreface{\@toklist={}
 \get@exampreface@body}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\endexampreface}
% This doesn't need to do anything.
%    \begin{macrocode}
\def\endexampreface{\relax}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\get@exampreface@body}
% This macro collects everything between it and the next |\end| token.  It then
% calls the |\find@exampreface@end| macro to check whether the |\end| token
% belongs to the \textsf{exampreface} environment or some other environment.
%    \begin{macrocode}
\long\def\get@exampreface@body#1\end{%
  \@toklist=\expandafter{\the\@toklist#1}%
  \find@exampreface@end}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\find@exampreface@end}
% If we are at |\end{exampreface}|, save everything to the appropriately named
% macro.  Otherwise, continue collecting tokens
%    \begin{macrocode}
\def\find@exampreface@end#1{%
  \def\@tempa{#1}%
  \ifx\@tempa\@Exampreface
    \expandafter\exampreface@wrap@up
  \else
    \@toklist=\expandafter{\the\@toklist\end{#1}}%
    \expandafter\get@exampreface@body
  \fi}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\exampreface@wrap@up}
% Save the scanned text to an appropriately named macro for later recall.
%    \begin{macrocode}
\def\exampreface@wrap@up{%
  \xdef\@examprefacetext{\the\@toklist}
  \end{exampreface}}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\@Exampreface}
% Store the name of the environment for comparison in |\find@exampreface@end|
%    \begin{macrocode}
\def\@Exampreface{exampreface}
%    \end{macrocode}
% \end{macro}
%
% \subsection{The \textsf{examclosing} environment}
%
% The \textsf{examclosing} environment behaves just like the \textsf{frontmatter}
% environment, except the enclosed text is saved in a differently-named
% macro.
%
% \begin{macro}{\examclosing}
% We define "\examclosing" so that it reads in and saves the surround text
% to a macro like in the \textsf{tabularx} environment.
%    \begin{macrocode}
\def\examclosing{\@toklist={}
 \get@examclosing@body}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\endexamclosing}
% This doesn't need to do anything.
%    \begin{macrocode}
\def\endexamclosing{\relax}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\get@examclosing@body}
% This macro collects everything between it and the next |\end| token.  It then
% calls the |\find@examclosing@end| macro to check whether the |\end| token
% belongs to the \textsf{examclosing} environment or some other environment.
%    \begin{macrocode}
\long\def\get@examclosing@body#1\end{%
  \@toklist=\expandafter{\the\@toklist#1}%
  \find@examclosing@end}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\find@examclosing@end}
% If we are at |\end{examclosing}|, save everything to the appropriately named
% macro.  Otherwise, continue collecting tokens
%    \begin{macrocode}
\def\find@examclosing@end#1{%
  \def\@tempa{#1}%
  \ifx\@tempa\@Examclosing
    \expandafter\examclosing@wrap@up
  \else
    \@toklist=\expandafter{\the\@toklist\end{#1}}%
    \expandafter\get@examclosing@body
  \fi}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\examclosing@wrap@up}
% Save the scanned text to an appropriately named macro for later recall.
%    \begin{macrocode}
\def\examclosing@wrap@up{%
  \xdef\@examclosingtext{\the\@toklist}
  \end{examclosing}}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\@Examclosing}
% Store the name of the environment for comparison in |\find@examclosing@end|
%    \begin{macrocode}
\def\@Examclosing{examclosing}
%    \end{macrocode}
% \end{macro}
%
% We now initialize "\@\examprefacetext" and "\@examclosingtext" to "\relax"
% so that in case the user doesn't specify any special text, nothing will be
% typeset in the actual exam.
%
%    \begin{macrocode}
\begin{exampreface}
  \relax
\end{exampreface}
%    \end{macrocode}
%
%    \begin{macrocode}
\begin{examclosing}
  \relax
\end{examclosing}
%    \end{macrocode}
%
% \subsection{The \textsf{keypreface} environment}
%
% The \textsf{keypreface} environment behaves just like the \textsf{frontmatter}
% environment, except the enclosed text is saved in a differently-named
% macro---one that will be printed just beneath any text specified in the
% \textsf{keytop} environment.
%
% \begin{macro}{\keypreface}
% We define "\keypreface" so that it reads in and saves the surround text
% to a macro like in the \textsf{tabularx} environment.
%    \begin{macrocode}
\def\keypreface{\@toklist={}
 \get@keypreface@body}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\endkeypreface}
% This doesn't need to do anything.
%    \begin{macrocode}
\def\endkeypreface{\relax}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\get@keypreface@body}
% This macro collects everything between it and the next |\end| token.  It then
% calls the |\find@keypreface@end| macro to check whether the |\end| token
% belongs to the \textsf{keypreface} environment or some other environment.
%    \begin{macrocode}
\long\def\get@keypreface@body#1\end{%
  \@toklist=\expandafter{\the\@toklist#1}%
  \find@keypreface@end}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\find@keypreface@end}
% If we are at "\end{keypreface}", save everything to the appropriately named
% macro.  Otherwise, continue collecting tokens
%    \begin{macrocode}
\def\find@keypreface@end#1{%
  \def\@tempa{#1}%
  \ifx\@tempa\@Keypreface
    \expandafter\keypreface@wrap@up
  \else
    \@toklist=\expandafter{\the\@toklist\end{#1}}%
    \expandafter\get@keypreface@body
  \fi}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\keypreface@wrap@up}
% Save the scanned text to an appropriately named macro for later recall.
%    \begin{macrocode}
\def\keypreface@wrap@up{%
  \xdef\@keyprefacetext{\the\@toklist}
  \end{keypreface}}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\@Keypreface}
% Store the name of the environment for comparison in "\find@keypreface@end"
%    \begin{macrocode}
\def\@Keypreface{keypreface}
%    \end{macrocode}
% \end{macro}
%
% \subsection{The \textsf{keyclosing} environment}
%
% The \textsf{keyclosing} environment behaves just like the \textsf{frontmatter}
% environment, except the enclosed text is saved in a differently-named
% macro.
%
% \begin{macro}{\keyclosing}
% We define "\keyclosing" so that it reads in and saves the surround text
% to a macro like in the \textsf{tabularx} environment.
%    \begin{macrocode}
\def\keyclosing{\@toklist={}
 \get@keyclosing@body}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\endkeyclosing}
% This doesn't need to do anything.
%    \begin{macrocode}
\def\endkeyclosing{\relax}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\get@keyclosing@body}
% This macro collects everything between it and the next "\end" token.  It then
% calls the "\find@keyclosing@end" macro to check whether the "\end" token
% belongs to the \textsf{keyclosing} environment or some other environment.
%    \begin{macrocode}
\long\def\get@keyclosing@body#1\end{%
  \@toklist=\expandafter{\the\@toklist#1}%
  \find@keyclosing@end}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\find@keyclosing@end}
% If we are at |\end{keyclosing}|, save everything to the appropriately named
% macro.  Otherwise, continue collecting tokens
%    \begin{macrocode}
\def\find@keyclosing@end#1{%
  \def\@tempa{#1}%
  \ifx\@tempa\@Keyclosing
    \expandafter\keyclosing@wrap@up
  \else
    \@toklist=\expandafter{\the\@toklist\end{#1}}%
    \expandafter\get@keyclosing@body
  \fi}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\keyclosing@wrap@up}
% Save the scanned text to an appropriately named macro for later recall.
%    \begin{macrocode}
\def\keyclosing@wrap@up{%
  \xdef\@keyclosingtext{\the\@toklist}
  \end{keyclosing}}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\@Keyclosing}
% Store the name of the environment for comparison in "\find@keyclosing@end"
%    \begin{macrocode}
\def\@Keyclosing{keyclosing}
%    \end{macrocode}
% \end{macro}
%
% We now initialize "\@keyprefacetext" and "\@keyclosingtext" to "\relax"
% so that in case the user doesn't specify any special text, nothing will be
% typeset in the actual exam.
%
%    \begin{macrocode}
\begin{keypreface}
  \relax
\end{keypreface}
%    \end{macrocode}
%
%    \begin{macrocode}
\begin{keyclosing}
  \relax
\end{keyclosing}
%    \end{macrocode}
%
% \subsection{Macros for manipulating token lists}
% In order to make manipulation of token lists easier, (or, really, to just
% increase the overall readability of the syntax, we define some macros
% to manipulate token lists.
%
% \begin{macro}{\append}
% |\append{|\meta{string}|}{|\meta{token-list}|}| appends \meta{string}
% to the rightmost end of \meta{token-list}.
%
% Note that we use |\@separator| between entries in the list, rather than
% |\\| as recommended in the \TeX{}book, since we will later want to redefine
% the list separator and \emph{execute} the list.  If we used |\\|, we run into
% potential conflicts in executing the list, since \LaTeXe, at any given point
% during the execution of the list entry, might redefine |\\| in such a way as to
% screw things up.
%    \begin{macrocode}
\def\append#1#2{%
  \global\temp@toks=\expandafter{\the #2\@separator {#1}}%
  \global#2=\expandafter{\the\temp@toks}}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\length}
% The usage is |\length{|\meta{token-list}|}|.  So |\length{\foo}| computes the
% length of the token list |\foo|, where it is assumed that the separator
% between list elements is |\@separator|.
%    \begin{macrocode}
\def\length#1{%
  \setcounter{length}{0}
  \def\@separator##1{\stepcounter{length}}\the#1}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\remove}
% |\remove{|\meta{number}|}{|\meta{token-list}|}| removes the \meta{number}th
% element from the token list \meta{token-list}, redefining \meta{token-list}
% to be that list.  As always, we assume that the list elements are separated
% by the token |\@separator|.
%    \begin{macrocode}
\def\remove#1#2{%
  \exam@c=0
  \temptoks={}
  \def\@separator##1{\advance\exam@c by 1\relax
     \ifnum\exam@c=#1\relax%
     \setcounter{extract}{##1}%
     \else\append{##1}{\temptoks}
     \fi}
  \the#2
  \global#2=\expandafter{\the\temptoks}}
%    \end{macrocode}
% \end{macro}
%
% \subsection{Scrambling the order of the questions}
%
% The general idea is that, for each section, we form a token list named
% |\qlist@sec|\meta{c} of the form:
% |{\@separator{1}\@separator{2}|\ldots|\@separator{|\meta{n}|}}|
% where \meta{c} is the number of the section (in roman numerals), and
% \meta{n} is the number of questions in section \meta{c}.  We call a while-loop
% to loop \meta{n} times, each time removing a randomly
%  chosen entry from the token list. This entry is appended to the list
% |\qlist@scr|\meta{m}@|sec|\meta{c}, where
% \meta{m} is the current value (in roman numerals) of the form we are
% generating. Once this list is built, we can then typeset the questions by
% simply redefining |\@separator| and executing the new scrambled list.
%
% \begin{macro}{\@enddocumenthook}
% Redefine the |\@enddocumenthook| so that the questions are automatically
% generated without the user needing to insert a command after all the question
% blocks.  This has the ironic effect that absolutely nothing the user sees in
% the input file causes the text to be created.  I haven't been real careful to
% prevent collisions occurring with other applications that use
% |\@enddocumenthook|, so this could be a source of problems.
%
% \medskip
%
% First, save any previously assigned meaning to
% |\@enddocumenthook| so that we don't clobber it.
%    \begin{macrocode}
\let\@oldenddocumenthook=\@enddocumenthook
\def\@enddocumenthook{%
  \global\let\pagebreak=\old@pagebreak
  \global\let\BreakPageOnVersion=\exam@DoneReadingBreakPageOnVersion
%    \end{macrocode}
% We need to create the number of forms specified, so we loop over the
% total number of forms.  Initialize the counter \texttt{version} to 1 (this
% is the value we will loop over) and step the counter \texttt{form} (so that 
% we actually generate all the forms that the user wants).
%    \begin{macrocode}
  \setcounter{version}{1} \stepcounter{form}
%    \end{macrocode}
% Loop until the value of \texttt{version} exceeds \texttt{form}.
%    \begin{macrocode}
  \@whilenum \value{version}<\value{form} \do{%
%    \end{macrocode}
% But two passes per exam are needed (once for the exam, once for the answers),
% so we create an ``inner-loop'' to take care of this problem.  Since we will
% want the answer key to have a slightly different format (where the
% answers are included!) it pays to make this on the second pass.
%    \begin{macrocode}
    \setcounter{innerloop}{0}
    \@whilenum \value{innerloop}<\value{passes} \do{%
%    \end{macrocode}
% Reset the page counter so that no new exams or answer keys start on page 17
% (for example).
%    \begin{macrocode}
        \setcounter{page}{1}
        \setcounter{questionnumber}{0}
        \setcounter{enumi}{0}
        \setcounter{equation}{0}
%    \end{macrocode}
% If |innerloop|=0, we are making an exam. If not, we are making an answer key,
% so take the appropriate action for the top of the form.  Since we won't want
% absent-minded professors to make 200+ copies of the answer key, print the
% title of the answer key using |\Huge| to call attention to it.
%    \begin{macrocode}
        \ifx\@frontmattertext\relax
          \relax
        \else
          \@frontmattertext \pagebreak
        \fi
        \ifnum \value{innerloop}=0 %
          \@examtopmatter
          \answerfalse
          \iftwocolumn \begin{multicols}{2}\fi
          \@examprefacetext
        \else%
          \@keytopmatter
          \answertrue
          \iftwocolumn \begin{multicols}{2}\fi
          \@keyprefacetext
        \fi
        \setcounter{sectionindex}{0}
%    \end{macrocode}
% And finally, once we have started making the exam, we need to loop over each
% individual section.
%    \begin{macrocode}
        \@whilenum \value{sectionindex}<\value{section} \do{%
            \addtocounter{sectionindex}{1}
%    \end{macrocode}
% Store the value of |sectionindex| in the macro
% |\current@section| so we can use it in |\csname|\ldots |\endcsname| pairs.
%    \begin{macrocode}
            \edef\current@section{\thesectionindex}
%    \end{macrocode}
% Compute the length of |\qlist@sec|\meta{current@section}
%    \begin{macrocode}
            \length{\csname qlist@sec\current@section\endcsname}
%    \end{macrocode}
% Here's the main meat of the code.  Pick a random number between 1 and the
% length of |\qlist@sec|\meta{current@section}.  Store it in the counter |\exam@m|.
% Remove the $m$th element from |\qlist@sec|\meta{current@section} and store it
% as the first element in |\qlist@scr@sec|\meta{current@section}.  Continuing to
% do this will rearrange the order of the questions.  The only thing we need to
% keep in mind is to change the requested range of the random number to keep in
% mind the fact that we removed an element from
% |\qlist@sec|\meta{current@section}.  Although there are faster ways to do
% this, I chose to opt for more readable code, since this part of the program is
% so complex anyway.
%    \begin{macrocode}
            \csname sec\current@section @fixed\endcsname
            \ifnotfixed
              \ifanswer \relax
              \else \typeout{Scrambling section \current@section, question: }\fi
              \exam@ScrambleQuestionList
            \else
              \csname qlist@scr@sec\current@section\endcsname=\expandafter{%
               \the\csname qlist@sec\current@section\endcsname}
            \fi
%    \end{macrocode}
% Finally, the code dealing with processing the different types of questions.  Each
% type of question is processed in the exact same way: we perform an |\ifx| to
% test to see what type of question it is, and if it is that type of question,
% execute the following code, which redefines |\begin{answer}|\ldots
% |\end{answer}| as necessary to achieve the desired effects.
%
% \medskip
%
% We now conduct a series of tests to see exactly what type of question the
% current section contains.  Once a match is made, the appropriate macro
% (defined below) is expanded, which typesets all the questions in that section.
%    \begin{macrocode}
        \csname RS@sec@\current@section\endcsname
        \@ifundefined{sec\current@section break}
          {\relax}
          {\pagebreak}
        \ifsection{\@matching}{\@matchingmain}
        \ifsection{\@shortanswer}{\@shortanswermain}
        \ifsection{\@fillin}{\@fillinmain}
        \ifsection{\@MC}{\@multiplechoicemain}
        \ifsection{\@truefalse}{\@truefalsemain}}
        \ifnum \value{innerloop}=0 %
          \@examclosingtext\par
        \else
          \@keyclosingtext\par
        \fi
        \iftwocolumn \end{multicols} \fi
        \ifx\@endmattertext\relax
          \relax
        \else
          \pagebreak \@endmattertext \par
        \fi
        \pagebreak
        \addtocounter{innerloop}{1}}
      \@OneKeyHook
%    \end{macrocode}
% This is the end of the code for the inner while-loop (that generates both
% exams and answer sheets. Toggle the value of "\ifanswer" so that
% on the second pass we generate answers to the questions.  Add one to
% "innerloop", too.
%    \begin{macrocode}
      \stepcounter{version}
      \setcounter{sectionindex}{0}
%    \end{macrocode}
% This is the end of the code for the form while-loop.  Add -1 to |form| to
% keep track of the number of copies we've made, and redefine the
% |\qlist@sec|\meta{section} lists, since they were eaten up by |\remove|
% during the randomization. (This is what the next loop does.)
%    \begin{macrocode}
      \@whilenum \value{sectionindex}<\value{section} \do{%
        \addtocounter{sectionindex}{1}
        \edef\current@section{\thesectionindex}
        \csname sec\current@section @fixed\endcsname
        \ifnotfixed
          \exam@RestoreSectionLists
        \else \notfixedtrue\fi}
      }}

 \def\exam@RestoreSectionLists{%
   \csname qlist@sec\current@section\endcsname=\expandafter{\the\csname
     qlist@scr@sec\current@section\endcsname}
   \csname qlist@scr@sec\current@section\endcsname={}}

\def\exam@ScrambleQuestionList{%
  \length{\csname qlist@sec\current@section\endcsname}%
  \@whilenum \value{length}>0 \do{%
    \setrannum{\exam@m}{1}{\value{length}}%
    \remove{\exam@m}{\csname qlist@sec\current@section\endcsname}%
    \edef\@TEMP{\theextract}%
    \ifnum \value{length}=1
      \message{\@TEMP.}%
    \else 
      \message{\@TEMP,}%
    \fi
    \expandafter\append\expandafter{\@TEMP}{\csname
      qlist@scr@sec\current@section\endcsname}
    \length{\csname qlist@sec\current@section\endcsname}}}
%    \end{macrocode}
% \end{macro}
%
% \section{The layout of the exam}
%
% \begin{macro}{\NumberOfVersions}
% A user-level command to specify the number of version to make. Originally
% initialized to 1.
%    \begin{macrocode}
\def\NumberOfVersions#1{\setcounter{form}{#1}}
 \NumberOfVersions{1}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\instructions}
% Often one wants to preface a block of questions with a series of instructions
% for the student.  Previous releases had defined "\instructions" to do this,
% but that command is obsolete as of version 1.02.  Thus, we define
% "\instructions" to be "\relax".
%    \begin{macrocode}
\def\instructions{\relax}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\classdata}
% This command saves its argument in the macro |\classdata| (it is intended for
% data like the class name and course number. E.g., Phil 29: Critical
% Reasoning).  Originally initialized to |\relax|.
%    \begin{macrocode}
\def\class#1{\def\classdata{#1}}
\def\classdata{\relax}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\examtype}
% Allows one to specify the name of the exam (midterm, final, etc.)
%    \begin{macrocode}
\def\examname#1{\def\examtype{#1}}
 \examname{Exam}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\namedata}
% Placed in a |\parbox|, |\namedata| contains all the standard fields
% ordinarily required in a large university setting.  If you should be so lucky as
% to teach at a place where student numbers are not required, simply delete the
% appropriate line.
%    \begin{macrocode}
\def\namedata{Name: \hrulefill \\[\namedata@vspace]
              Student Number: \hrulefill \\[\namedata@vspace]
              TA: \hrulefill \\[\namedata@vspace]
              Date: \hrulefill}
%    \end{macrocode}
% \end{macro}
%
% \subsection{Types of questions}
%
% Five types of questions have special
% formats: true/false, fill-in-the-blank, short answer, multiple choice, and
% fixed (meaning: don't swap the order of the questions).  Since we will want to
% distinguish between these types of questions when creating the exam, we need
% a way to identify a particular section with a particular type of question.
% We do this in the following way: for each section, we define a macro
% |\type@sec|\meta{section} consisting of a single string containing the name
% of the type of question.  That string will later be used to identify the 
% section as such and take appropriate action.
%
% First, some necessary definitions.
%    \begin{macrocode}
\define@key{section}{title}[\relax]{%
  \expandafter\gdef\csname head@sec\thesection\endcsname{#1}}

\define@key{section}{resetcounter}[yes]{%
  \def\exam@arg{#1}%
  \ifx\exam@arg\exam@c@yes
    \expandafter
    \gdef\csname RS@sec@\thesection\endcsname{\resetcountertrue}
  \else
    \expandafter
    \gdef\csname RS@sec@\thesection\endcsname{\resetcounterfalse}
  \fi}

\define@key{section}{suppressprefix}[yes]{%
  \def\exam@arg{#1}%
  \ifx\exam@arg\exam@c@yes
    \expandafter
    \gdef\csname SP@sec@\thesection\endcsname{\resetcountertrue}
  \fi}

\define@key{section}{rearrange}[yes]{%
  \def\exam@arg{#1}%
  \ifx\exam@arg\exam@c@yes
    \exam@SectionIsNotFixed
  \else
    \exam@SectionIsFixed
  \fi}

\define@key{section}{keycolumns}[1]{%
  \ifnum#1=1\relax\else
    \ifx\@currentsectionname\@matching
    \message{Column specification ignored for the `matching' environment.}
    \else
    \exam@SetKeyColumns{#1}
    \fi
  \fi}
\define@key{section}{examcolumns}[1]{%
  \ifnum#1=1\relax\else
    \ifx\@currentsectionname\@matching
    \message{Column specification ignored for the `message' environment.}
    \else
    \exam@SetExamColumns{#1}
    \fi
  \fi}

\def\exam@definesectiontype#1{%
  \expandafter\gdef\csname type@sec\thesection\endcsname{#1}}
%    \end{macrocode}
% These next three macros just makes some following code easier to read.
%    \begin{macrocode}
\def\exam@SetKeyColumns#1{%
  \expandafter\gdef\csname sec\thesection @KCol\endcsname{#1}}
\def\exam@SetExamColumns#1{%
  \expandafter\gdef\csname sec\thesection @ECol\endcsname{#1}}
\def\exam@SectionIsNotFixed{%
   \expandafter\gdef\csname sec\thesection @fixed\endcsname{\notfixedtrue}}
\def\exam@SectionIsFixed{%
   \expandafter\gdef\csname sec\thesection @fixed\endcsname{\notfixedfalse}}
\def\exam@SetDefaultSectionTitle#1{
   \expandafter\gdef\csname head@sec\thesection\endcsname{#1}}
\def\exam@SetDefaultRecountState#1{
   \expandafter\gdef\csname RS@sec@\thesection\endcsname{#1}}
\def\exam@notitle{\exam@notitle}
%    \end{macrocode}
%
% \begin{environment}{matching}
% The environment for matching questions.
%    \begin{macrocode}
\newenvironment{matching}[1][]
  {\begin{section}
   \def\@currentsectionname{matching}
   \exam@definesectiontype{matching}
   \exam@SetDefaultRearrangeBehavior
   \exam@SetDefaultSectionTitle{\exam@notitle}
   \exam@SetDefaultRecountState{\exam@DefaultNumberingBehavior}
   \setkeys{section}{#1}%
   \setcounter{temp}{0}%
   \exam@GetInstructionsForMatchingSection}
  {\end{section}}

\long\def\exam@GetInstructionsForMatchingSection#1\pair{%
  \@ifundefined{sec\thesection q\thequestion break}
    {\@toklist={}}
    {\@toklist={\old@pagebreak}}
  \expandafter
  \xdef\csname inst@sec\thesection\endcsname{\the\@toklist#1}\pair}

\def\exam@MatchingLineLength{.5in}
\def\exam@MatchingLineAfterSkip{6pt}
\def\pair#1#2{%
  \expandafter
  \xdef\csname sec\thesection @\thequestion @tag\endcsname{\thetemp}
  \expandafter
  \gdef\csname sec\thesection @qst\thequestion\endcsname{%
    \exam@PairA{\item[]\hskip-\leftmargin 
      \ifanswer
      \underline{\makebox[\exam@MatchingLineLength][c]{\exam@GetMatchingAnswer}}%
      \hskip\exam@MatchingLineAfterSkip #1
      \else
      \rule{\exam@MatchingLineLength}{.4pt}%
      \hskip\exam@MatchingLineAfterSkip #1
      \fi}}
  \stepcounter{question}
  \expandafter
  \xdef\csname sec\thesection @\thequestion @tag\endcsname{\thetemp}
  \expandafter
  \gdef\csname sec\thesection @qst\thequestion\endcsname{%
    \exam@PairB{\item\relax\exam@SaveMatchingAnswer #2}}
  \stepcounter{temp}
  \stepcounter{question}}

\def\exam@GetMatchingAnswer{%
  \edef\exam@tag{\csname sec\current@section @\current@question
  @tag\endcsname}%
  \csname matching@answer@\exam@tag\endcsname}
\def\exam@SaveMatchingAnswer{%
  \edef\exam@tag{\csname sec\current@section @\current@question
    @tag\endcsname}%
  \expandafter
  \xdef\csname matching@answer@\exam@tag\endcsname{(\alph{enumi})}}

%    \end{macrocode}
% \end{environment}
%
% \begin{environment}{fillin}
% The environment for fill-in-the-blank questions.
%    \begin{macrocode}
\newenvironment{fillin}[1][]
  {\begin{section}
   \def\@currentsectionname{fillin}
   \exam@definesectiontype{fillin}
   \exam@SetDefaultRearrangeBehavior
   \exam@SetDefaultSectionTitle{\exam@notitle}
   \exam@SetDefaultRecountState{\exam@DefaultNumberingBehavior}
   \setkeys{section}{#1}%
   \get@instructions \relax}
  {\end{section}}
%    \end{macrocode}
% \end{environment}
%
%
% \begin{environment}{truefalse}
% The environment for true/false questions
%    \begin{macrocode}
\newenvironment{truefalse}[1][]
  {\begin{section}
   \def\@currentsectionname{truefalse}
   \exam@definesectiontype{truefalse}
   \exam@SetDefaultRearrangeBehavior
   \exam@SetDefaultSectionTitle{\exam@notitle}
   \exam@SetDefaultRecountState{\exam@DefaultNumberingBehavior}
   \setkeys{section}{#1}%
   \get@instructions \relax}
  {\end{section}}
%    \end{macrocode}
% \end{environment}
%
%
% \begin{environment}{shortanswer}
% The environment for short answer questions
%    \begin{macrocode}
\newenvironment{shortanswer}[1][]
  {\begin{section}
   \def\@currentsectionname{shortanswer}
   \exam@definesectiontype{shortanswer}
   \exam@SetDefaultRearrangeBehavior
   \exam@SetDefaultSectionTitle{\exam@notitle}
   \exam@SetDefaultRecountState{\exam@DefaultNumberingBehavior}   
   \setkeys{section}{#1}%
   \get@instructions \relax}
  {\end{section}}
%    \end{macrocode}
% \end{environment}
%
%
% \begin{environment}{multiplechoice}
% The environment for multiple choice questions
%    \begin{macrocode}
\newenvironment{multiplechoice}[1][]
  {\begin{section}
   \def\@currentsectionname{multiplechoice}
   \exam@definesectiontype{multiplechoice}
   \exam@SetDefaultRearrangeBehavior
   \exam@SetDefaultSectionTitle{\exam@notitle}
   \exam@SetDefaultRecountState{\exam@DefaultNumberingBehavior}   
   \setkeys{section}{#1}%
   \get@instructions \relax}
  {\end{section}}
%    \end{macrocode}
% \end{environment}
%
%
% \subsection{Loose Ends}
%
% The following macros just contain the environment names for identification in
% the main loop.
%    \begin{macrocode}
\def\@matching{matching}
\def\@fillin{fillin}
\def\@shortanswer{shortanswer}
\def\@truefalse{truefalse}
\def\@MC{multiplechoice}
%    \end{macrocode}
%
% \begin{macro}{\choice}
% This is one of the complicated macros in the package.  From the user's point
% of view, it indicates one answer to a multiple choice question.  Internally,
% |\choice| has to recognize whether it is the first or last question, taking
% different action depending on which it is.  Moreover, |\choice| has to format
% the label differently depending on whether or not we are creating an exam or a
% test. Note that, for reasons of sanity, I do not attempt to rearrange the order
% of the choices.
%    \begin{macrocode}
\newcommand{\exam@DefaultChoice}[2][]{%
  \ifchoice \relax
    \else \@mcstartlist \choicetrue
  \fi
  \ifanswer
    \if#1!%
      \ifbox \vspace{-\fboxsep}\item[\hfill\fbox{(\alph{choice})}\hfill] #2
        \vspace{-\fboxsep}%
      \else
         \item[\hfill\MCfont{(\alph{choice})}\hfill] \MCfont{#2}%
      \fi
    \else \item[\hfill(\alph{choice})\hfill] #2
    \fi 
    \stepcounter{choice}%
  \else
    \item[\hfill(\alph{choice})\hfill] #2
    \stepcounter{choice}%
  \fi}
\def\exam@EndChoiceListIfNecessary{%
  \ifchoice\end{list}\choicefalse\fi\setcounter{choice}{1}}
\let\choice=\exam@DefaultChoice
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\@mcstartlist}
% The parameters used to start the |\begin{list}| environment used in
% the |\choice| macro.
%    \begin{macrocode}
\def\@mcstartlist{\begin{list}{}{%
 \setlength{\topsep}{6pt}
 \setlength{\rightmargin}{0pt}
 \setlength{\labelwidth}{.35in}
 \setlength{\leftmargin}{.5in}
 \setlength{\labelsep}{.15in}
 \setlength{\itemindent}{0in}
 \setlength{\itemsep}{0pt}}}
%    \end{macrocode}
% \end{macro}
%
%
% \subsection{More User-Level Macros}
%
% \begin{macro}{\Fullpages}
% Just a command so that people can automatically set the page margins to
% full-sized pages.  Only defined for US letter, sorry.
%    \begin{macrocode}
\newcommand{\Fullpages}{\setlength{\topmargin}{0pc}
                        \setlength{\oddsidemargin}{0pc}
                        \setlength{\evensidemargin}{0pc}
                        \setlength{\headheight}{0pc}
                        \setlength{\headsep}{0pc}
                        \setlength{\textwidth}{6.5in}
                        \setlength{\textheight}{9in}}
%</othercode>
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\ConstantBlanks}
% Sets the blanks in \textsf{fillin} to be of constant width
%    \begin{macrocode}
\def\ConstantBlanks#1{\setlength{\blank@length}{#1}\cblankstrue\pblanksfalse}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\ProportionalBlanks}
% Sets the blanks in \textsf{fillin} to be proportional to the actual typeset
% width of the word.
%    \begin{macrocode}
\def\ProportionalBlanks#1{\edef\blank@multiplier{#1}\cblanksfalse\pblankstrue}
 \ProportionalBlanks{1}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\StudentInfoLineSpacing}
% Sets the amount of spacing between the lines in the |\namedata| macro.
%    \begin{macrocode}
\newcommand{\StudentInfoLineSpacing}[1]{\setlength{\namedata@vspace}{#1}}
 \StudentInfoLineSpacing{6pt}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\examtop}
% We define "\examtop" so that it reads in and saves the surround text
% to a macro like in the \textsf{tabularx} environment.
%    \begin{macrocode}
\def\examtop{\@toklist={}
 \get@examtop@body}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\endexamtop}
% This doesn't need to do much.
%    \begin{macrocode}
\def\endexamtop{\relax}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\get@examtop@body}
% This macro collects everything between it and the next |\end| token.  It then
% calls the |\find@examtop@end| macro to check whether the |\end| token
% belongs to the \textsf{question} environment or some other environment.
%    \begin{macrocode}
\long\def\get@examtop@body#1\end{%
  \@toklist=\expandafter{\the\@toklist#1}%
  \find@examtop@end}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\find@examtop@end}
% If we are at |\end{examtop}|, save everything to the appropriately named
% macro.  Otherwise, continue collecting tokens
%    \begin{macrocode}
\def\find@examtop@end#1{%
  \def\@tempa{#1}%
   \ifx\@tempa\@XMtop\expandafter\XMtop@wrap@up
    \else\@toklist=\expandafter{\the\@toklist\end{#1}}%
     \expandafter\get@examtop@body\fi}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\XMtop@wrap@up}
% Save the gathered information to a macro for later use.
%    \begin{macrocode}
\def\XMtop@wrap@up{\xdef\@examtopmatter{\the\@toklist}%
 \end{examtop}}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\@XMtop}
% Store the name of the environment for comparison in |\find@question@end|
%    \begin{macrocode}
\def\@XMtop{examtop}
%    \end{macrocode}
% \end{macro}
%
% \DescribeEnv{examtop}
% This environment is defined via the macros below.  It allows the user
% to customize the text at the top of the exam.  The example here is the
% default definition.
%    \begin{macrocode}
\begin{examtop}
\@@line{\parbox{3in}{\classdata \\
        \examtype, Form:~\fbox{\textsf{\Alph{version}}}}
        \hfill
        \parbox{3in}{\normalsize \namedata}}
        \bigskip
\end{examtop}
%    \end{macrocode}
%
% \begin{macro}{\keytop}
% We define "\keytop" so that it reads in and saves the surround text
% to a macro like in the \textsf{tabularx} environment.
%    \begin{macrocode}
\def\keytop{\@toklist={}
 \get@keytop@body}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\endkeytop}
% This doesn't need to do much.
%    \begin{macrocode}
\def\endkeytop{\relax}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\get@keytop@body}
% This macro collects everything between it and the next |\end| token.  It then
% calls the |\find@keytop@end| macro to check whether the |\end| token
% belongs to the \textsf{question} environment or some other environment.
%    \begin{macrocode}
\long\def\get@keytop@body#1\end{\@toklist=\expandafter{\the\@toklist#1}%
  \find@keytop@end}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\find@keytop@end}
% If we are at |\end{keytop}|, save everything to the appropriately named
% macro.  Otherwise, continue collecting tokens
%    \begin{macrocode}
\def\find@keytop@end#1{%
  \def\@tempa{#1}%
   \ifx\@tempa\@KYtop\expandafter\KYtop@wrap@up
    \else\@toklist=\expandafter{\the\@toklist\end{#1}}%
     \expandafter\get@keytop@body\fi}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\KYtop@wrap@up}
% Save the gathered information to a macro for later use.
%    \begin{macrocode}
\def\KYtop@wrap@up{\xdef\@keytopmatter{\the\@toklist}%
 \end{keytop}}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\@KYtop}
% Store the name of the environment for comparison in |\find@keytop@end|
%    \begin{macrocode}
\def\@KYtop{keytop}
%    \end{macrocode}
% \end{macro}
%
% \DescribeEnv{keytop}
% This environment is defined via the macros below.  It allows the user
% to customize the text at the top of the exam.  The example here is the
% default definition.
%    \begin{macrocode}
\begin{keytop}
\@@line{\hfill \Huge Answer Key
        for Exam \fbox{\textsf{\Alph{version}}}\hfill}
        \bigskip
\end{keytop}
%    \end{macrocode}
%
% \MakeShortVerb{\"}
% \DeleteShortVerb{\|}
% \begin{macro}{\word}
% This macro enables one to insert minor (read: inessential) changes in wording
% between different versions of the test.  For example, including the
% command "\word|{Carl Sagan} {Bob Dole}|" in the text will cause the phrase
% ``Carl Sagan'' to be printed in that place on every odd-numbered test and ``Bob Dole''
% to be printed in that place on every even-numbered test.  In the event that
% more versions of the test are requested than options given, "\word" acts as if
% the given list of options repeats itself.  I.e., in the above case, if we requested
% four different versions of the examn, "\word" will behave as if the
% command given were "\word|{Carl Sagan} {Bob Dole} {Carl Sagan} {Bob Dole}|".
%
% \smallskip
%
% First we define a string to be equal to a space so that we can allow for any
% number of spaces to be given in the option list.
%    \begin{macrocode}
\def\@Space{ }
%    \end{macrocode}
% The "\word" macro just saves the option-list in a macro named "\@wordlist" and
% invokes the macro "\@wordmain" which most of the work.
%    \begin{macrocode}
\def\word#1{\def\@wordlist{#1}\expandafter\@wordmain\@wordlist \stop}

\def\@wordmain#1{%
%    \end{macrocode}
% Check to see if the current argument is a space.  If so, ignore it.
%    \begin{macrocode}
  \ifx#1\@Space
   \let\@next=\@wordmain
  \else
%    \end{macrocode}
% If the current argument is not a space, but is "\stop", this means that
% fewer options were given than versions of the test were requested.  Thus we
% need to start "\@wordmain" again at the beginning of the option list.  This
% is what the "\@dumpthenstart" macro does.
%    \begin{macrocode}
   \ifx#1\stop
    \let\@next=\@dumpthenstart
   \else%
%    \end{macrocode}
% If the current argument is none of the above, then it is an actual option.
% However, we must check whether it is the option we want.  If it is, then
% typeset it and gobble everything up to (and including) "\stop".  If not, just
% throw the current argument away and get the next one.
%    \begin{macrocode}
    \ifnum\value{wordtemp}=\value{version}%
     #1\let\@next=\@gobbletostop
    \else
     \stepcounter{wordtemp}%
     \let\@next=\@wordmain
    \fi
   \fi
  \fi
  \@next}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\@dumpthenstart}
%    \begin{macrocode}
\def\@dumpthenstart{\expandafter\@wordmain\@wordlist \stop}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\@gobbletostop}
%    \begin{macrocode}
\def\@gobbletostop#1\stop{\relax\setcounter{wordtemp}{1}}
%    \end{macrocode}
% \end{macro}
%
% \section{Macros to format types of questions}
%
% \begin{macro}{\ifsection}
% This macro increases the readability of the code in the definition
% of "\@enddocumenthook" when we test to see just which type of questions
% the current section consists of.
%    \begin{macrocode}
\def\ifsection#1#2{%
 \expandafter\ifx\csname type@sec\current@section\endcsname#1#2\fi}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\exam@AddSectionTitle}
% Typeset the section title, if there is one; otherwise, don't do anything.
%    \begin{macrocode}
\def\exam@AddSectionTitle{%
  \expandafter\ifx\csname head@sec\current@section\endcsname\exam@notitle
    \relax
  \else
    \par\vspace{\beforesectsep}
    \noindent {\exam@sectionfont \exam@sectionprefix
      \csname head@sec\current@section\endcsname}
    \par\vspace{\aftersectsep}
  \fi}


%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\exam@AddSectionInstructions}
% Add instructions for the section, if any were give; otherwise, don't
% do anything.
%    \begin{macrocode}
\def\exam@AddSectionInstructions{%
  \ifx\csname inst@sec\current@section\endcsname\relax
    \relax
  \else
    \vspace{\beforeinstsep}
    \csname inst@sec\current@section\endcsname
    \par\vspace{\afterinstsep}
  \fi}
\let\exam@AddSectionInstructionsHook=\exam@AddSectionInstructions
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\exam@MakeAnswerEnvironGobble}
% This macro redefines the \textsf{answer} environment so that it
% ignores (``gobbles'') everything inside it.  Used when typesetting
% the exam.
%    \begin{macrocode}
\def\exam@MakeAnswerEnvironGobble{%
  \def\answer{\get@noanswer@body}
  \def\endanswer{\relax}
  \def\@S{answer}
  \long\def\get@noanswer@body##1\end{\find@answer@end}
  \def\find@answer@end##1{\def\@tempa{##1}
    \ifx\@tempa\@S\expandafter\finish@up
    \else\expandafter\get@noanswer@body\fi}
  \def\finish@up{\end{answer}}}
%    \end{macrocode}
% \end{macro}
%
%
% \begin{macro}{\exam@TypesetQuestions}
% With the exception of the \textsf{truefalse} environment, 
% all of the other kinds of questions are handled in the sam
% way.  (The \textsf{truefalse} environment is special because of
% the wide blanks it places at the left margin.)
%    \begin{macrocode}
\def\exam@PossiblyStartMultipleColumnsInKey{%
  \ifanswer
    \@ifundefined{sec\current@section @KCol}
      \relax
      {\begin{multicols}{\csname sec\current@section @KCol\endcsname}}
  \fi}
\def\exam@PossiblyEndMultipleColumnsInKey{%
  \ifanswer
    \@ifundefined{sec\current@section @KCol}
      \relax
      {\end{multicols}}
  \fi}
\def\exam@PossiblyStartMultipleColumnsInExam{%
  \ifanswer\relax\else
    \@ifundefined{sec\current@section @ECol}
      \relax
      {\begin{multicols}{\csname sec\current@section @ECol\endcsname}}
  \fi}
\def\exam@PossiblyEndMultipleColumnsInExam{%
  \ifanswer\relax\else
    \@ifundefined{sec\current@section @ECol}
      \relax
      {\end{multicols}}
  \fi}

\def\exam@EnumerateStyle{1.}
\def\exam@DefaultEnumerateStyle{1.}

\def\enumerate{%
  \ifnum \@enumdepth >3 \@toodeep\else
     \advance\@enumdepth \@ne
     \edef\@enumctr{enum\romannumeral\the\@enumdepth}\fi
  \@ifnextchar[{\exam@@enum@}{\@enum@}}
\def\exam@@enum@[#1]{\expandafter\@@enum@\expandafter[#1]}

\def\exam@TypesetQuestions{%
  \exam@PossiblyStartMultipleColumnsInKey
  \exam@PossiblyStartMultipleColumnsInExam
  \ifx\exam@EnumerateStyle\exam@DefaultEnumerateStyle
    \begin{enumerate}
  \else
    \begin{enumerate}[\exam@EnumerateStyle]
  \fi
    \ifresetcounter
      \setcounter{enumi}{0}
    \else
%      \message{QuestionNumber: \thequestionnumber}%
      \setcounter{enumi}{\value{questionnumber}}
    \fi
    \the\csname qlist@scr@sec\current@section\endcsname
    \setcounter{questionnumber}{\value{enumi}}
  \end{enumerate}
  \exam@PossiblyEndMultipleColumnsInKey
  \exam@PossiblyEndMultipleColumnsInExam}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\@shortanswermain}
%
% Define |\@separator| in the appropriate fashion for this kind of question.
%    \begin{macrocode}
\def\@shortanswermain{%
  \def\question{%
    \item 
    \@beforequestionhook
    \exam@SAGobbleQuestionHook}
  \def\endquestion{\@endquestionhook}
  \def\@separator##1{\setcounter{temp}{##1}
    \edef\current@question{\thetemp}
    \csname sec\current@section @qst\current@question\endcsname}
%    \end{macrocode}
% If we are makingf a answer, set up the \textsf{answer} environment so that
% it prints the answer along with an identifier.
%    \begin{macrocode}
  \ifanswer
    \exam@DefineAnswerEnvironment
  \else
    \exam@MakeAnswerEnvironGobble
  \fi
  \exam@AddSectionTitle
  \exam@AddSectionInstructionsHook
  \exam@TypesetQuestions}
%
% \begin{macro}{\@matchingmain}
%
% Define |\@separator| in the appropriate fashion for this kind of question.
%    \begin{macrocode}
\def\@matchingmain{%
  \def\@separator##1{\setcounter{temp}{##1}
    \edef\current@question{\thetemp}
    \csname sec\current@section @qst\current@question\endcsname}
%    \end{macrocode}
% If we are makingf a answer, set up the \textsf{answer} environment so that
% it prints the answer along with an identifier.
%    \begin{macrocode}
  \exam@AddSectionTitle
  \exam@AddSectionInstructionsHook
  \def\exam@PairA##1{##1}
  \def\exam@PairB##1{}
  \edef\exam@questionnumber{\thequestionnumber}%
  \begin{multicols}{2}\raggedcolumns
  \exam@TypesetQuestions
  \def\exam@PairA##1{}
  \def\exam@PairB##1{##1}
  \def\exam@EnumerateStyle{(a)}
  \columnbreak
  \exam@TypesetQuestions
  \end{multicols}
  \setcounter{enumi}{\exam@questionnumber}%
  \setcounter{questionnumber}{\exam@questionnumber}}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\@fillinmain}
% As with all of the section routines, we need to
% define |\@separator| again.
%    \begin{macrocode}
\def\@fillinmain{%
  \def\question{%
    \item 
    \@beforequestionhook
    \exam@FillinBlockGobbleQuestionHook}
  \def\endquestion{\@endquestionhook}
  \def\@separator##1{\setcounter{temp}{##1}
    \edef\current@question{\thetemp}
    \csname sec\current@section @qst\current@question\endcsname}
%    \end{macrocode}
% Define |\blank| so that it either makes just an underline or an underline
% with the word printed over it.
%    \begin{macrocode}
  \ifanswer 
    \ifcblanks
    \def\blank##1{\mbox{\rule[-1pt]{\blank@length}{.4pt}%
        \hspace{-\blank@length}%
        \makebox[\blank@length]{##1}}}
    \fi
    \ifpblanks
    \def\blank##1{\settowidth{\temp@length}{##1}%
      \mbox{\rule[-1pt]{\blank@multiplier\temp@length}{.4pt}%
        \hspace{-\blank@multiplier\temp@length}%
        \makebox[\blank@multiplier\temp@length]{##1}}}
    \fi
  \else
    \ifcblanks
      \def\blank##1{\rule[-1pt]{\blank@length}{.4pt}}
    \fi
    \ifpblanks
      \def\blank##1{\settowidth{\temp@length}{##1}%
        \rule[-1pt]{\blank@multiplier\temp@length}{.4pt}}
    \fi
  \fi
  \exam@AddSectionTitle
  \exam@AddSectionInstructionsHook
  \exam@TypesetQuestions}
%    \end{macrocode}
% \end{macro}
%
%
% \begin{macro}{\@multiplechoicemain}
% This contains all the commands to be invoked if the current section in
% the main loop is a multiple-choice environment.
%    \begin{macrocode}
\def\@multiplechoicemain{\setcounter{mcchoice}{1}
%    \end{macrocode}
% First, redefine the "\question", "\endquestion" pairs in case someone
% used the \textsf{block} environment within multiplechoice.
%    \begin{macrocode}
   \def\question{\@mcitem
     \@beforequestionhook
     \exam@MCGobbleQuestionHook}
   \def\endquestion{\@endquestionhook
     \exam@EndChoiceListIfNecessary\@mcenditem}
%    \end{macrocode}
% Now define "\@separator" so that it calls the appropriate macro.
%    \begin{macrocode}
   \def\@separator##1{\setcounter{temp}{##1}
     \edef\current@question{\thetemp}
     \csname sec\current@section @qst\current@question\endcsname}
   \exam@AddSectionTitle
   \exam@AddSectionInstructionsHook
   \exam@TypesetQuestions}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\@truefalsemain}
% This contains all the commands to be invoked if the current section in the
% main loop is a true/false environment.
%    \begin{macrocode}
\def\@truefalsemain{%
%    \end{macrocode}
% First, define "\@separator" so that it calls the macro where the right
% question is saved.
%    \begin{macrocode}
   \def\@separator##1{\setcounter{temp}{##1}
     \edef\current@question{\thetemp}
     \csname sec\current@section @qst\current@question\endcsname}
%    \end{macrocode}
% Redefine "\question" and "\endquestion" to work properly in case someone
% used a \textsf{block} environment to keep a couple of true/false questions
% together
%    \begin{macrocode}
   \def\question{%
     \edef\current@question{\thetemp.\theexam@tf}%
     \exam@TFItem \@beforequestionhook}
   \def\endquestion{\stepcounter{exam@tf}\@endquestionhook}
%    \end{macrocode}
% Since the appearance of a true/false question depends heavily on whether we
% are typesetting an answer key or not, we have to make a number of definitions
% (or redefinitions) depending upon whether we are currently making an answer
% key.
%    \begin{macrocode}
     \ifanswer
%    \end{macrocode}
% If we are typesetting an answer, we don't want the argument given to
% "\answer" to do anything, so we define it so that it gobbles its argument.
%    \begin{macrocode}
       \def\answer##1{\relax}
%    \end{macrocode}
% If we aren't typesetting an answer, then we are typesetting a test. We
% must define "\answer" so that it saves the answer in
% a macro so that it can be recalled when the answer key is typeset.
%    \begin{macrocode}
     \else
       \def\answer##1{\expandafter\gdef
          \csname tf@sol\current@question\endcsname{##1}}
          \vspace{-\smallskipamount}
     \fi
     \exam@AddSectionTitle
     \exam@AddSectionInstructionsHook
     \exam@StartTrueFalseList
     \the\csname qlist@scr@sec\current@section\endcsname
     \end{list}}

\def\exam@StartTrueFalseList{%
  \ifanswer
  \begin{list}{}
    {\setlength{\rightmargin}{0pt}
      \setlength{\labelwidth}{.4in}
      \setlength{\leftmargin}{.5in}
      \setlength{\labelsep}{.1in}
      \setlength{\itemindent}{0in}
      \setlength{\itemsep}{3pt}}
%    \the\csname qlist@scr@sec\current@section\endcsname
%  \end{list}
  \else
  \begin{list}%{$\underline{\mbox{\hphantom{false}}}$}
    {\setlength{\rightmargin}{0pt}
      \setlength{\labelwidth}{.35in}
      \setlength{\leftmargin}{.5in}
      \setlength{\labelsep}{.15in}
      \setlength{\itemindent}{0in}
      \setlength{\itemsep}{3pt}}
%    \the\csname qlist@scr@sec\current@section\endcsname
%  \end{list}
  \fi}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\@endquestionhook}
% This provides a ``hook'' for inserting something (like a place to put points)
% at the end of every question.  Originally initialized to just "\relax"
%    \begin{macrocode}
\def\@endquestionhook{\relax}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\@beforequestionhook}
% This provides a ``hook'' for inserting something (like a place to put points)
% at the start of every question.  Originally initialized to just "\relax".
%    \begin{macrocode}
\def\@beforequestionhook{\relax}
%    \end{macrocode}
% \end{macro}
%
%\def\@endquestionhook{\par\hfill Score: \fbox{\phantom{M}}}
%
% \begin{macro}{\exam@DefaultTFItem}
% This is a fancy version of "\item" that must be used for the true/false
% environment because of the way the exams and answers are printed.
%    \begin{macrocode}
\def\exam@DefaultTFItem{%
  \ifanswer
    \item[\uline{\makebox[.5in][c]{\csname tf@sol\current@question\endcsname}}]
  \else
    \item[\uline{\makebox[.5in][r]{\hss}}]
  \fi}
\let\exam@TFItem=\exam@DefaultTFItem
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\@mcitem}
% This is a fancy version of "\item" that must be used for the multiple
% choice environment because we want to put the question text inside a
% minipage to keep it from being broken across pages.
%    \begin{macrocode}
\def\@mcitem{\item\begin{minipage}[t]{\linewidth}}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\@mcenditem}
% This completes the minipage environment begun above.
%    \begin{macrocode}
\def\@mcenditem{\end{minipage}}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\answerspace}
% Inserts a particular amount of space for an answer to be given only on
% the actual exam.  The space is suppressed in the answer key.
%    \begin{macrocode}
\def\answerspace#1{\ifanswer \relax \else \vspace*{#1} \fi}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\exam}
% Whatever is given as an argument to this macro will only be included if we
% are typesetting an exam.
%    \begin{macrocode}
\def\exam#1{\ifanswer\relax\else #1\fi}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\key}
% Whatever is given as an argument to this macro will only be included if we
% are typesetting an answer key.
%    \begin{macrocode}
\def\key#1{\ifanswer#1\fi}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\examvspace}
% Insert the specified amount of vertical space only if we are making an
% exam. At the same time, check to see if the command was a starred command or
% not.
%    \begin{macrocode}
\def\examvspace{\@ifnextchar*{\@examvspacestar}{\@examvspace}}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\@examvspacestar}
%    \begin{macrocode}
\def\@examvspacestar*#1{\exam{\vspace*{#1}}}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\@examvspace}
%    \begin{macrocode}
\def\@examvspace#1{\exam{\vspace{#1}}}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\examhspace}
% Insert the specified amount of horizontal space only if we are making an
% exam. At the same time, check to see if the command was a starred command or
% not.
%    \begin{macrocode}
\def\examhspace{\@ifnextchar*{\@examhspacestar}{\@examhspace}}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\@examhspacestar}
%    \begin{macrocode}
\def\@examhspacestar*#1{\exam{\hspace*{#1}}}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\@examhspace}
%    \begin{macrocode}
\def\@examhspace#1{\exam{\hspace{#1}}}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\keyvspace}
% Insert the specified amount of vertical space only if we are making an
% exam. At the same time, check to see if the command was a starred command or
% not.
%    \begin{macrocode}
\def\keyvspace{\@ifnextchar*{\@keyvspacestar}{\@keyvspace}}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\@keyvspacestar}
%    \begin{macrocode}
\def\@keyvspacestar*#1{\key{\vspace*{#1}}}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\@keyvspace}
%    \begin{macrocode}
\def\@keyvspace#1{\key{\vspace{#1}}}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\keyhspace}
% Insert the specified amount of vertical space only if we are making an
% exam. At the same time, check to see if the command was a starred command or
% not.
%    \begin{macrocode}
\def\keyhspace{\@ifnextchar*{\@keyhspacestar}{\@keyhspace}}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\@keyhspacestar}
%    \begin{macrocode}
\def\@keyhspacestar*#1{\key{\hspace*{#1}}}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\@keyhspace}
%    \begin{macrocode}
\def\@keyhspace#1{\key{\hspace{#1}}}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\label}
% The multiple processing of exams generates error messages if we use
% the default definition of |\label|.  The following hack lets one
% use |\label| inside of exams with the expected effect.
%    \begin{macrocode}
\let\exam@oldlabel=\label
\renewcommand{\label}[1]{\exam{\exam@oldlabel{\theversion @#1}}}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\ref}
% And a similar hack to get |\ref| working as well\ldots
%    \begin{macrocode}
\let\exam@oldref=\ref
\renewcommand{\ref}[1]{\exam@oldref{\theversion @#1}}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\problem}
% Occasionally one needs to put several different numbered questions
% inside of a single block (I didn't first encounter this until I wrote
% my first scantron exam).  This macro gives a way to do this consistently
% with the numbering scheme for the rest of the section.
%    \begin{macrocode}
\def\problem{\stepcounter{enumi}\theenumi}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\SectionFont}
% Here's an easy way to change the font used in section headings.
%    \begin{macrocode}
\def\SectionFont#1{\def\exam@sectionfont{#1}}
\SectionFont{\large} % The default value
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\NoRearrange}
% The default behavior is to rearrange the questions.  This macro changes
% it.
%    \begin{macrocode}
\let\exam@SetDefaultRearrangeBehavior=\exam@SectionIsNotFixed
\def\NoRearrange{%
  \let\exam@SetDefaultRearrangeBehavior=\exam@SectionIsFixed}
%    \end{macrocode}
% \end{macro}
%
%    \begin{macrocode}
\def\exam@resetcounter{%
  \ifresetcounter
    \setcounter{enumi}{0}
    \global\resetcounterfalse
  \else
    \setcounter{enumi}{\value{questionnumber}}
  \fi}
%    \end{macrocode}
%
% \begin{macro}{\SectionPrefix}
% This macro let's one change the appearance of the section number
% which preceeds the title.
%    \begin{macrocode}
\def\SectionPrefix#1{%
  \def\exam@sectionprefix{%
    \@ifundefined{SP@sec@\current@section}
      {#1}
      \relax}}
\SectionPrefix{Section \arabic{sectionindex}. \space}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\ContinuousNumbering}
% When called, changes the default numbering behavior across sections.
%    \begin{macrocode}
\def\ContinuousNumbering{%
  \let\exam@DefaultNumberingBehavior=\resetcounterfalse}
\let\exam@DefaultNumberingBehavior=\resetcountertrue
%    \end{macrocode}
% \end{macro}
%

% \begin{macro}{\ShortKey}
% This macro only includes the answers in the key.
%    \begin{macrocode}
\def\exam@SAGobbleQuestionHook{\relax}% default values
\def\exam@MCGobbleQuestionHook{\relax}
\def\exam@TFGobbleQuestionHook{\relax}
\def\exam@FillinGobbleQuestionHook{\relax}
\def\exam@FillinBlockGobbleQuestionHook{\relax}
\def\exam@TrueFalseGobbleBlockInstructionsHook{\relax}
\def\exam@answer{answer}
\def\exam@enumerate{enumerate}
\def\exam@list{list}

\long\def\exam@BlockGobbleToBeginEnumerate#1\begin{%
  \exam@CheckNextTokenForEnumerate}
\def\exam@CheckNextTokenForEnumerate#1{%
  \def\@tempa{#1}%
  \ifx\@tempa\exam@enumerate
    \exam@BeginEnumerate
  \else
    \expandafter\exam@BlockGobbleToBeginEnumerate
  \fi}
\def\exam@BeginEnumerate{\begin{enumerate}}

\long\def\exam@TrueFalseBlockInstructionsGobble#1\exam@StartTrueFalseList{%
  \exam@StartTrueFalseList}

\long\def\exam@GobbleUpToAnswer#1\begin{%
  \exam@CheckNextTokenForAnswer}
\def\exam@CheckNextTokenForAnswer#1{%
  \def\@tempa{#1}%
  \ifx\@tempa\exam@answer
    \let\exam@next=\exam@PrepareToTypesetAnswer
  \else
    \ifx\@tempa\@endquestionhook
      \let\exam@next=\@endquestionhook
    \else
      \let\exam@next=\exam@GobbleUpToAnswer
  \fi\fi\exam@next}
\def\exam@PrepareToTypesetAnswer{\begin{answer}}

\long\def\exam@GobbleUpToEndQuestionHook#1\@endquestionhook{%
  \@endquestionhook}

\long\def\exam@GobbleUpToChoice#1\choice{\choice}

\def\exam@FillinGobbleNext#1{%
  \ifx #1\blank
    \let\exam@next=\exam@FillinInsertBlank
  \else
    \ifx #1\@endquestionhook
      \def\exam@insertedblank{no}%
      .\let\exam@next=\relax
    \else
      \let\exam@next=\exam@FillinGobbleNext
  \fi\fi\exam@next}

\def\exam@FillinBlockGobbleNext#1{%
  \ifx #1\blank
    \let\exam@next=\exam@FillinBlockInsertBlank
  \else
    \ifx #1\end
      \let\exam@next=\exam@CheckNextTokenForQuestion
    \else
      \let\exam@next=\exam@FillinBlockGobbleNext
  \fi\fi\exam@next}
\def\exam@CheckNextTokenForQuestion#1{%
  \def\@tempa{#1}%
  \ifx\@tempa\@Q
    \expandafter\exam@FillinWrapupAndEndQuestion
  \else
    \expandafter\exam@FillinBlockGobbleNext
  \fi}
\def\exam@FillinWrapupAndEndQuestion{.\end{question}}

\def\exam@insertedblank{no}
\def\exam@FillinInsertBlank#1{%
  \ifx\exam@insertedblank\exam@c@no
    #1\def\exam@insertedblank{yes}%
  \else
    , #1\relax
  \fi\exam@FillinGobbleNext}
\def\exam@FillinBlockInsertBlank#1{%
  \ifx\exam@insertedblank\exam@c@no
    #1\def\exam@insertedblank{yes}%
  \else
    , #1\relax
  \fi\exam@FillinBlockGobbleNext}

\def\ShortKey{%
  \def\exam@AddSectionInstructionsHook{%
    \ifanswer\relax\else\exam@AddSectionInstructions\fi}
  \def\exam@SAGobbleQuestionHook{%
    \ifanswer\expandafter\exam@GobbleUpToAnswer\fi}
  \def\exam@MCGobbleQuestionHook{%
    \ifanswer\expandafter\exam@GobbleUpToChoice\fi}
  \def\exam@FillinGobbleQuestionHook{%
    \ifanswer\expandafter\exam@FillinGobbleNext\fi}
  \def\exam@FillinBlockGobbleQuestionHook{%
    \ifanswer\expandafter\exam@FillinBlockGobbleNext\fi}
  \def\exam@GobbleBlockInstructionsHook{%
    \ifanswer\expandafter\exam@BlockGobbleToBeginEnumerate\fi}
  \def\exam@TrueFalseGobbleBlockInstructionsHook{%
    \ifanswer\expandafter\exam@TrueFalseBlockInstructionsGobble\fi}
  \let\choice=\exam@ShortChoice}

\def\exam@ShortChoice{%
  \ifanswer
    \expandafter\exam@ShortKeyChoice
  \else
    \expandafter\exam@DefaultChoice
  \fi}

\newif\ifOneCorrectAnswerAlreadyGiven
\OneCorrectAnswerAlreadyGivenfalse
\def\MultipleChoiceShortKeyPrefix#1{%
  \def\exam@MultipleChoiceShortKeyPrefix{#1}}
\MultipleChoiceShortKeyPrefix{}

\newcommand{\exam@ShortKeyChoice}[2][]{%
  \if#1!%
    \ifOneCorrectAnswerAlreadyGiven
    , (\alph{choice})
    \else
    \exam@MultipleChoiceShortKeyPrefix
    (\alph{choice})%
    \OneCorrectAnswerAlreadyGiventrue
    \fi
   \else
     \stepcounter{choice}%
  \fi\ignorespaces}

\def\exam@ShortTFItem{%
  \ifanswer
    \item[\csname tf@sol\current@question\endcsname]
  \else
    \item
  \fi}

\def\exam@c@yes{yes}
\def\exam@c@no{no}

\def\exam@MCMinipageWrapperBegin{\begin{minipage}[t]{\noexpand\linewidth}}
\def\exam@MCMinipageWrapperEnd{\end{minipage}}

\def\DefineAnswerWrapper#1#2{%
  \def\exam@DefineAnswerEnvironment{%
    \def\answer{#1}\def\endanswer{#2}}}
\DefineAnswerWrapper{\begin{description}\item[Answer:]}{\end{description}}

\let\old@pagebreak=\pagebreak
\def\exam@none{none}
\def\pagebreak{%
  \ifx\@currentsectionname\exam@none
    \setcounter{temp}{\thesection}
    \stepcounter{temp}
    \expandafter\def\csname sec\thetemp break\endcsname{\old@pagebreak}
  \else
    \expandafter
    \gdef\csname sec\thesection q\thequestion break\endcsname{\old@pagebreak}
  \fi}

\def\exam@WhileReadingBreakPageOnVersion#1{%
  \ifx\@currentsectionname\exam@none
    \setcounter{temp}{\thesection}%
    \stepcounter{temp}%
    \expandafter\def\csname sec\thetemp break\endcsname{%
      \ifnum\value{version}=#1\relax
        \old@pagebreak
      \fi}
  \else
    \expandafter
    \gdef\csname sec\thesection q\thequestion break\endcsname{%
      \ifnum\c@version=#1
        \old@pagebreak
      \fi}
  \fi}
\def\exam@DoneReadingBreakPageOnVersion#1{%
  \ifnum\value{version}=#1\relax
    \old@pagebreak
    \fi}
\let\BreakPageOnVersion=\exam@WhileReadingBreakPageOnVersion

\def\@currentsectionname{none}

\def\IncludeFromFile#1{\def\exam@includefile{#1}}

\def\InsertChunk#1{%
  \def\exam@chunklookup{#1}%
  \input{\exam@includefile}}

\def\chunk#1{%
  \def\@tempa{#1}%
  \ifx\@tempa\exam@chunklookup
    \relax
  \else
    \expandafter\exam@GobbleRestOfChunk
  \fi}
\def\endchunk{\relax}

\def\exam@chunk{chunk}
\long\def\exam@GobbleRestOfChunk#1\end{%
  \exam@CheckNextTokenForChunk}
\def\exam@CheckNextTokenForChunk#1{%
  \def\@tempa{#1}%
  \ifx\@tempa\exam@chunk
    \end{chunk}%
  \else
    \expandafter\exam@GobbleRestOfChunk
  \fi}


%    \end{macrocode}
% \end{macro}