\chapter{Conclusion}
\section{Objectives Achieved}
The CAL tool {\rm newogg} was written to enhance {\rm oggin} to make it
an informative and effective teaching aid by changing the format and
enriching the contents of its output.  {\rm newogg} displays
picture-like output on the screen after each instruction cycle.  Besides
the contents of the components of the cpu, {\rm newogg} also displays
memory locations accessed during the instruction cycle; the previous,
the current and the next instructions in assembly language mnemonics;
and an explanation as well as comments for each instruction.  Moreover,
locations concerned are highlighted when significant changes occur or
particular attention is required.  Thus the output results are more
convenient and impressive for examining the operation of instructions,
and the teaching results are more informative and effective.  All those
objectives have been achieved by {\rm newogg}, implementing the target
of this project.  
\\
\\
Furthermore, the {\rm newogg} also results in an unexpected side-effect
that it could be used as a debugging package for {\rm oggin} program.
However, to do this requires a method of assembling {\rm oggin} assembly
language instructions into {\rm oggin} machine code data, see the
following section.
\section{Possible Future Development}
\subsection{Adding the Assembler}
{\rm newogg} is only an interpreter and disassembler.  It can be
extended to be a complete {\rm oggin} machine by adding the assembler
part to it.  Thus the input data can be assembly language mnemonics.
These are three ways of implementation : \\
\begin{enumerate}
\item  Join {\rm newogg} to the original {\rm oggin} which, after the assembler
part, can be revised to have two choices for different formats of
output.  The user can select the original interpreter for a copy of
one-line output or select {\rm newogg} for demonstrating the results on
the screen one by one.
\item  Write an assembler and add to {\rm newogg} and make it become an
independent new virtual machine.
\item  If the {\rm oggin} program can be used only as an assembler,
without it continuing into the interpreter part, then it could be used
to assemble a file for {\rm newogg}, either by sending it through a UNIX
pipe, or using a temporary intermediate file.
\end{enumerate}
\subsection{Extensions of the Output}
\subsubsection{Creating a Window for Stack}
{\rm newogg} was designed to display its output on a vt100 terminal which has a
standard screen size of 23 lines and 80 columns.  For some instructions, a
screen of this size
is not big enough to show all the necessary information.  With the output of
the instruction {\it trp}, for instance, the two fixed addresses (10 and 11
octal) accessed for fetching the address of the trap service routine are not
shown on the
screen.  What is highlighted in the main memory window is the two stack
locations accessed, where the pc and psw have been pushed.
In other words, the main memory window is showing the
stack, ignoring the accessed locations at addresses 10 and 11 octal.  If 
{\rm newogg} is run on a system with a larger screen display, such as a Hewlett
Packard Workstation with a screen size of 46 lines and 128 columns, another
window could be created to show the stack, including the top forty
locations of memory, independently.  As a result, the stack is separated
from the main memory obviously and all the accessed locations, covered by
the main memory or in the stack, can be displayed and highlighted at the
same time.
\\
\\
\noindent
However, showing the stack and the main memory in the same window
emphasizes that they are both in the same physical memory.
\subsubsection{Grouping the Binary Codes}
Binary codes of each instruction are not displayed in the format of three
digits in one group, which would be easier to match with each
corresponding octal digit.  Somebody could revise {\rm newogg} and group
the binary digits into six groups.  In a system using a bigger screen,
the binary values of acc, source and operation result in the comm\_win,
can also be revised and displayed in the grouped format.  It would be
more considerate to the user if corresponding octal codes of those
binary values are displayed side by side.
\subsection{Implementing Extra Opcodes}
The full {\rm oggin} instruction set has three unused opcodes in group 1
and 2080 unused opcodes in group 2, (which could be further subdivided
into more groups) see appendix C.  If an attempt is
make to execute one of these instructions, {\rm newogg} generates an
error message.  In real processors in real computers, undefined
instructions can produce undefined effects.  Some processors just ignore
the instruction, and some make an exception call to allow the programmer
to handle them.  The way {\rm newogg} handles those is a bit like an
exception call, generating an error and stopping the program.
\\
\\
\noindent
To add extra instructions to the {\rm newogg} interpreter, you just need
to add the required extra selections to the {\it case} statements in
{\it newg1.c} and {\it newg2.c} to decode and execute them, and add the
extra mnemonics in {\it newm.c} and {\it glovar.h}. (The six files
which make up {\rm newogg} are installed in /usr/fsb/wlw/project on {\rm
crown} and /user/csplw1/project on {\rm forth}.)