\documentstyle[12pt]{article}
\topmargin=-0.5in
\oddsidemargin=-0.225in
\textwidth=6.5in
\textheight=9in
\newcommand{\I}[1]{{\tt #1}}
\setlength{\parskip}{1em}
\setlength{\parindent}{0in}
\begin{document}

\begin{center}\bf
        {\Large Com Sci 221 --- Programming Languages}\\[3ex]
        {\Large\bf Recording Your CS221 Work on Scheme, ML, Prolog Interpreters}\\[3ex]
        (by Steph Bailey, revised by Mike O'Donnell, Dec. 1993)
\end{center}

For assignments done with interpreters (the Scheme, ML, and Prolog
processors used in this course are interpreters), the best
presentation of the computations is usually a {\em typescript\/}
containing a listing of your program, a listing of your test file, and
all the output generated by the execution of your test file (see below
on how to create a typescript).  Once logged in, you will create your
source program with the text editor of your choice.  If you don't know
how to use a Unix text editor, type: {\tt emacs} at the shell prompt,
and then type the back-space key followed by {\tt t}, and this will
start you on a tutorial about how to use emacs, a popular text editor.

After you have created and debugged your program (more on that below),
you must create the {\em typescript\/} mentioned above.  Before you
can create a typescript, you must create a testing file for your
program which contains {\em exactly what you would type to ML to run
the test cases for your program.} The reason why you should use a test
file, rather than just typing the test cases directly in at the
keyboard is that using the {\em delete\/} key (a very common
occurrence with computers) causes the transcript file to become a
horrible, illisible mess.  If the commands used to create your
typescript come from a file, this problem won't arise.

The test file should cause your program to be read (compiled) in,
and run exciting and illustrative test cases, thus
really impressing me with how well your program performs.  The test file
should also contain occasional {\em output\/} statements which print a
running commentary on what is happening in your test cases, since the
contents of the test file are not printed as they are executed.
Otherwise I may have a difficult time figuring out which commands
produced what output, and I will get annoyed.

As an example, I'll pretend that I'm handing
in an ML program called {\tt prop.sml}.  My test data file is called
{\tt mltestdata}.

\begin{verbatim}
(* Comments are enclosed in these things, but they don't print out *)
use "prop.sml";   (* Load the program *)
output std_out "\n*** finding cnf(OR(AND(VAR \"p\",VAR \"q\"), VAR \"r\"));\n";
cnf(OR(AND(VAR "p",VAR "q"), VAR "r"));  (* Boring test case! *)
(* Now set the magic system variable to enable printing the whole list *)
System.Control.Print.printDepth := 10;    (* This can be useful *)
output std_out "\n*** finding cnf(OR(AND(VAR \"p\",VAR \"q\"), VAR \"r\"));\n";
cnf(OR(AND(VAR "p",VAR "q"), VAR "r"));  (* I wouldn't get a good grade
                                            for these tests *)
\end{verbatim}

After you have created a test data file, you may create the script
with the following sequence of commands:
\begin{itemize}
\item {\tt script} at the shell prompt.  This
will start a new shell where everything which appears on your terminal
will be recorded.
\item {\tt cat progfile} where progfile is the name of your program
file (repeat this command several times if you have several source
files).
\item {\tt cat testfile} where testfile is the name of your testing
file.
\item {\tt ml <testfile} this invokes ML to execute all the statements
that you have in your testfile.  It will complete all the commands and
then exit back to the shell
\item {\tt exit}  this exits the script shell and closes the
typescript file.
\end{itemize}

\end{document}