Computer Science and Compiler Design An

Introduction with C 1st Edition by Compiler

Generators in C ISBN pdf download

https://ebookball.com/product/computer-science-and-compiler-
design-an-introduction-with-c-1st-edition-by-compiler-generators-
in-c-isbn-11720/

Explore and download more ebooks or textbooks

at ebookball.com
 Get Your Digital Files Instantly: PDF, ePub, MOBI and More
Quick Digital Downloads: PDF, ePub, MOBI and Other Formats

Data Structures Algorithms and Applications in C With Microsoft

Compiler 2nd Edition by Sartaj Sahni ISBN 0929306325 9780929306322

https://ebookball.com/product/data-structures-algorithms-and-
applications-in-c-with-microsoft-compiler-2nd-edition-by-sartaj-
sahni-isbn-0929306325-9780929306322-15764/

Data Structures Algorithms and Applications in C++ With Microsoft

Compiler 1st edition by Sartaj Sahni ISBN 007236226X 978-0072362268

https://ebookball.com/product/data-structures-algorithms-and-
applications-in-c-with-microsoft-compiler-1st-edition-by-sartaj-
sahni-isbn-007236226x-978-0072362268-16394/

Data Structures Algorithms and Applications in C++ With Microsoft

Compiler 2nd edition by Michael Goodrich, Roberto Tamassia, David
Mount ISBN B005FHM6X2 978-0470383278

https://ebookball.com/product/data-structures-algorithms-and-
applications-in-c-with-microsoft-compiler-2nd-edition-by-michael-
goodrich-roberto-tamassia-david-mount-
isbn-b005fhm6x2-978-0470383278-16544/

Modern Compiler Design Worldwide Series in Computer Science 1st

Edition by Dick Grune, Henri Bal, Ceriel Jacobs, Koen Langendoen ISBN
0471976970 9780471976974

https://ebookball.com/product/modern-compiler-design-worldwide-
series-in-computer-science-1st-edition-by-dick-grune-henri-bal-
ceriel-jacobs-koen-langendoen-
isbn-0471976970-9780471976974-19834/
An Introduction to Design Patterns in C with Qt 4 1st Edition by Alan
Ezust, Paul Ezust ISBN 0131879057 9780131879058

https://ebookball.com/product/an-introduction-to-design-patterns-
in-c-with-qt-4-1st-edition-by-alan-ezust-paul-ezust-
isbn-0131879057-9780131879058-12664/

Introduction to Design Patterns in C with Qt 2nd Edition by Alan

Ezust, Paul Ezust ISBN 0132851636 9780132851633

https://ebookball.com/product/introduction-to-design-patterns-in-
c-with-qt-2nd-edition-by-alan-ezust-paul-ezust-
isbn-0132851636-9780132851633-14144/

Compiler Construction: An Advanced Course 1st Edition by Bauer FL ISBN

3540069585 9783540069584

https://ebookball.com/product/compiler-construction-an-advanced-
course-1st-edition-by-bauer-fl-
isbn-3540069585-9783540069584-13264/

Developing Software for Symbian OS An Introduction to c in C++ 1st

edition by Steve Babin ISBN 0470018453 978-0470018453

https://ebookball.com/product/developing-software-for-symbian-os-
an-introduction-to-c-in-c-1st-edition-by-steve-babin-
isbn-0470018453-978-0470018453-13366/

Algorithms for Compiler Design 1st edition by Kakde ISBN ‎

1584501006 978-1584501008

https://ebookball.com/product/algorithms-for-compiler-design-1st-
edition-by-kakde-isbn-aeurz-1584501006-978-1584501008-17748/
 Compilers and Compiler Generators
an introduction with C++
© P.D. Terry, Rhodes University, 1996

p.terry@ru.ac.za

This is a set of Adobe PDF® files of the text of my book "Compilers and Compiler Generators - an
introduction with C++", published in 1997 by International Thomson Computer Press. The original
edition is now out of print, and the copyright has reverted to me.

The book is also available in other formats. The latest versions of the distribution and details of
how to download up-to-date compressed versions of the text and its supporting software and
courseware can be found at http://www.scifac.ru.ac.za/compilers/

The text of the book is Copyright © PD Terry. Although you are free to make use of the material
for academic purposes, the material may not be redistributed without my knowledge or permission.

File List
The 18 chapters of the book are filed as chap01.pdf through chap18.pdf
The 4 appendices to the book are filed as appa.pdf through appd.pdf
The original appendix A of the book is filed as appa0.pdf
The contents of the book is filed as contents.pdf
The preface of the book is filed as preface.pdf
An index for the book is filed as index.pdf. Currently (January 2000) the page numbers refer
to an A4 version in PCL® format available at
http://www.scifac.ru.ac.za/compilers/longpcl.zip. However, software tools like GhostView
may be used to search the files for specific text.
The bibliography for the book is filed as biblio.pdf

Change List
18-October-1999 - Pre-release
12-November-1999 - First official on-line release
16-January-2000 - First release of Postscript version (incorporates minor corrections to
chapter 12)
17-January-2000 - First release of PDF version
 Compilers and Compiler Generators © P.D. Terry, 2000

PREFACE
This book has been written to support a practically oriented course in programming language
translation for senior undergraduates in Computer Science. More specifically, it is aimed at students
who are probably quite competent in the art of imperative programming (for example, in C++,
Pascal, or Modula-2), but whose mathematics may be a little weak; students who require only a
solid introduction to the subject, so as to provide them with insight into areas of language design
and implementation, rather than a deluge of theory which they will probably never use again;
students who will enjoy fairly extensive case studies of translators for the sorts of languages with
which they are most familiar; students who need to be made aware of compiler writing tools, and to
come to appreciate and know how to use them. It will hopefully also appeal to a certain class of
hobbyist who wishes to know more about how translators work.

The reader is expected to have a good knowledge of programming in an imperative language and,
preferably, a knowledge of data structures. The book is practically oriented, and the reader who
cannot read and write code will have difficulty following quite a lot of the discussion. However, it
is difficult to imagine that students taking courses in compiler construction will not have that sort of
background!

There are several excellent books already extant in this field. What is intended to distinguish this
one from the others is that it attempts to mix theory and practice in a disciplined way, introducing
the use of attribute grammars and compiler writing tools, at the same time giving a highly practical
and pragmatic development of translators of only moderate size, yet large enough to provide
considerable challenge in the many exercises that are suggested.

Overview

The book starts with a fairly simple overview of the translation process, of the constituent parts of a
compiler, and of the concepts of porting and bootstrapping compilers. This is followed by a chapter
on machine architecture and machine emulation, as later case studies make extensive use of code
generation for emulated machines, a very common strategy in introductory courses. The next
chapter introduces the student to the notions of regular expressions, grammars, BNF and EBNF,
and the value of being able to specify languages concisely and accurately.

Two chapters follow that discuss simple features of assembler language, accompanied by the
development of an assembler/interpreter system which allows not only for very simple assembly,
but also for conditional assembly, macro-assembly, error detection, and so on. Complete code for
such an assembler is presented in a highly modularized form, but with deliberate scope left for
extensions, ranging from the trivial to the extensive.

Three chapters follow on formal syntax theory, parsing, and the manual construction of scanners
and parsers. The usual classifications of grammars and restrictions on practical grammars are
discussed in some detail. The material on parsing is kept to a fairly simple level, but with a
thorough discussion of the necessary conditions for LL(1) parsing. The parsing method treated in
most detail is the method of recursive descent, as is found in many Pascal compilers; LR parsing is
only briefly discussed.
The next chapter is on syntax directed translation, and stresses to the reader the importance and
usefulness of being able to start from a context-free grammar, adding attributes and actions that
allow for the manual or mechanical construction of a program that will handle the system that it
defines. Obvious applications come from the field of translators, but applications in other areas
such as simple database design are also used and suggested.

The next two chapters give a thorough introduction to the use of Coco/R, a compiler generator
based on L- attributed grammars. Besides a discussion of Cocol, the specification language for this
tool, several in-depth case studies are presented, and the reader is given some indication of how
parser generators are themselves constructed.

The next two chapters discuss the construction of a recursive descent compiler for a simple
Pascal-like source language, using both hand-crafted and machine-generated techniques. The
compiler produces pseudo-code for a hypothetical stack-based computer (for which an interpreter
was developed in an earlier chapter). "On the fly" code generation is discussed, as well as the use of
intermediate tree construction.

The last chapters extend the simple language (and its compiler) to allow for procedures and
functions, demonstrate the usual stack-frame approach to storage management, and go on to discuss
the implementation of simple concurrent programming. At all times the student can see how these
are handled by the compiler/interpreter system, which slowly grows in complexity and usefulness
until the final product enables the development of quite sophisticated programs.

The text abounds with suggestions for further exploration, and includes references to more
advanced texts where these can be followed up. Wherever it seems appropriate the opportunity is
taken to make the reader more aware of the strong and weak points in topical imperative languages.
Examples are drawn from several languages, such as Pascal, Modula-2, Oberon, C, C++, Edison
and Ada.

Support software

An earlier version of this text, published by Addison-Wesley in 1986, used Pascal throughout as a
development tool. By that stage Modula-2 had emerged as a language far better suited to serious
programming. A number of discerning teachers and programmers adopted it enthusiastically, and
the material in the present book was originally and successfully developed in Modula-2. More
recently, and especially in the USA, one has witnessed the spectacular rise in popularity of C++,
and so as to reflect this trend, this has been adopted as the main language used in the present text.
Although offering much of value to skilled practitioners, C++ is a complex language. As the aim of
the text is not to focus on intricate C++programming, but compiler construction, the supporting
software has been written to be as clear and as simple as possible. Besides the C++ code, complete
source for all the case studies has also been provided on an accompanying IBM-PC compatible
diskette in Turbo Pascal and Modula-2, so that readers who are proficient programmers in those
languages but only have a reading knowledge of C++ should be able to use the material very
successfully.

Appendix A gives instructions for unpacking the software provided on the diskette and installing it
on a reader’s computer. In the same appendix will be found the addresses of various sites on the
Internet where this software (and other freely available compiler construction software) can be
found in various formats. The software provided on the diskette includes
 Emulators for the two virtual machines described in Chapter 4 (one of these is a simple
accumulator based machine, the other is a simple stack based machine).

The one- and two-pass assemblers for the accumulator based machine, discussed in Chapter 6.

A macro assembler for the accumulator-based machine, discussed in Chapter 7.

Three executable versions of the Coco/R compiler generator used in the text and described in
detail in Chapter 12, along with the frame files that it needs. (The three versions produce
Turbo Pascal, Modula-2 or C/C++ compilers)

Complete source code for hand-crafted versions of each of the versions of the Clang compiler
that is developed in a layered way in Chapters 14 through 18. This highly modularized code
comes with an "on the fly" code generator, and also with an alternative code generator that
builds and then walks a tree representation of the intermediate code.

Cocol grammars and support modules for the numerous case studies throughout the book that
use Coco/R. These include grammars for each of the versions of the Clang compiler.

A program for investigating the construction of minimal perfect hash functions (as discussed
in Chapter 14).

A simple demonstration of an LR parser (as discussed in Chapter 10).

Use as a course text

The book can be used for courses of various lengths. By choosing a selection of topics it could be
used on courses as short as 5-6 weeks (say 15-20 hours of lectures and 6 lab sessions). It could also
be used to support longer and more intensive courses. In our university, selected parts of the
material have been successfully used for several years in a course of about 35 - 40 hours of lectures
with strictly controlled and structured, related laboratory work, given to students in a pre-Honours
year. During that time the course has evolved significantly, from one in which theory and formal
specification played a very low key, to the present stage where students have come to appreciate the
use of specification and syntax-directed compiler-writing systems as very powerful and useful tools
in their armoury.

It is hoped that instructors can select material from the text so as to suit courses tailored to their
own interests, and to their students’ capabilities. The core of the theoretical material is to be found
in Chapters 1, 2, 5, 8, 9, 10 and 11, and it is suggested that this material should form part of any
course based on the book. Restricting the selection of material to those chapters would deny the
student the very important opportunity to see the material in practice, and at least a partial selection
of the material in the practically oriented chapters should be studied. However, that part of the
material in Chapter 4 on the accumulator-based machine, and Chapters 6 and 7 on writing
assemblers for this machine could be omitted without any loss of continuity. The development of
the small Clang compiler in Chapters 14 through 18 is handled in a way that allows for the later
sections of Chapter 15, and for Chapters 16 through 18 to be omitted if time is short. A very wide
variety of laboratory exercises can be selected from those suggested as exercises, providing the
students with both a challenge, and a feeling of satisfaction when they rise to meet that challenge.
Several of these exercises are based on the idea of developing a small compiler for a language
similar to the one discussed in detail in the text. Development of such a compiler could rely entirely
on traditional hand-crafted techniques, or could rely entirely on a tool-based approach (both
approaches have been successfully used at our university). If a hand-crafted approach were used,
Chapters 12 and 13 could be omitted; Chapter 12 is largely a reference manual in any event, and
could be left to the students to study for themselves as the need arose. Similarly, Chapter 3 falls into
the category of background reading.

At our university we have also used an extended version of the Clang compiler as developed in the
text (one incorporating several of the extensions suggested as exercises) as a system for students to
study concurrent programming per se, and although it is a little limited, it is more than adequate for
the purpose. We have also used a slightly extended version of the assembler program very
successfully as our primary tool for introducing students to the craft of programming at the
assembler level.

Limitations

It is, perhaps, worth a slight digression to point out some things which the book does not claim to
be, and to justify some of the decisions made in the selection of material.

In the first place, while it is hoped that it will serve as a useful foundation for students who are
already considerably more advanced, a primary aim has been to make the material as accessible as
possible to students with a fairly limited background, to enhance the background, and to make them
somewhat more critical of it. In many cases this background is still Pascal based; increasingly it is
tending to become C++ based. Both of these languages have become rather large and complex, and
I have found that many students have a very superficial idea of how they really fit together. After a
course such as this one, many of the pieces of the language jigsaw fit together rather better.

When introducing the use of compiler writing tools, one might follow the many authors who
espouse the classic lex/yacc approach. However, there are now a number of excellent LL(1) based
tools, and these have the advantage that the code which is produced is close to that which might be
hand-crafted; at the same time, recursive descent parsing, besides being fairly intuitive, is powerful
enough to handle very usable languages.

That the languages used in case studies and their translators are relative toys cannot be denied. The
Clang language of later chapters, for example, supports only integer variables and simple
one-dimensional arrays of these, and has concurrent features allowing little beyond the simulation
of some simple textbook examples. The text is not intended to be a comprehensive treatise on
systems programming in general, just on certain selected topics in that area, and so very little is said
about native machine code generation and optimization, linkers and loaders, the interaction and
relationship with an operating system, and so on. These decisions were all taken deliberately, to
keep the material readily understandable and as machine-independent as possible. The systems may
be toys, but they are very usable toys! Of course the book is then open to the criticism that many of
the more difficult topics in translation (such as code generation and optimization) are effectively
not covered at all, and that the student may be deluded into thinking that these areas do not exist.
This is not entirely true; the careful reader will find most of these topics mentioned somewhere.

Good teachers will always want to put something of their own into a course, regardless of the
quality of the prescribed textbook. I have found that a useful (though at times highly dangerous)
technique is deliberately not to give the best solutions to a problem in a class discussion, with the
optimistic aim that students can be persuaded to "discover" them for themselves, and even gain a
sense of achievement in so doing. When applied to a book the technique is particularly dangerous,
but I have tried to exploit it on several occasions, even though it may give the impression that the
author is ignorant.

Another dangerous strategy is to give too much away, especially in a book like this aimed at
courses where, so far as I am aware, the traditional approach requires that students make far more
of the design decisions for themselves than my approach seems to allow them. Many of the books
in the field do not show enough of how something is actually done: the bridge between what they
give and what the student is required to produce is in excess of what is reasonable for a course
which is only part of a general curriculum. I have tried to compensate by suggesting what I hope is
a very wide range of searching exercises. The solutions to some of these are well known, and
available in the literature. Again, the decision to omit explicit references was deliberate (perhaps
dangerously so). Teachers often have to find some way of persuading the students to search the
literature for themselves, and this is not done by simply opening the journal at the right page for
them.

Acknowledgements
I am conscious of my gratitude to many people for their help and inspiration while this book has
been developed.

Like many others, I am grateful to Niklaus Wirth, whose programming languages and whose
writings on the subject of compiler construction and language design refute the modern trend
towards ever-increasing complexity in these areas, and serve as outstanding models of the way in
which progress should be made.

This project could not have been completed without the help of Hanspeter Mössenböck (author of
the original Coco/R compiler generator) and Francisco Arzu (who ported it to C++), who not only
commented on parts of the text, but also willingly gave permission for their software to be
distributed with the book. My thanks are similarly due to Richard Cichelli for granting permission
to distribute (with the software for Chapter 14) a program based on one he wrote for computing
minimal perfect hash functions, and to Christopher Cockburn for permission to include his
description of tonic sol-fa (used in Chapter 13).

I am grateful to Volker Pohlers for help with the port of Coco/R to Turbo Pascal, and to Dave
Gillespie for developing p2c, a most useful program for converting Modula-2 and Pascal code to
C/C++.

I am deeply indebted to my colleagues Peter Clayton, George Wells and Peter Wentworth for many
hours of discussion and fruitful suggestions. John Washbrook carefully reviewed the manuscript,
and made many useful suggestions for its improvement. Shaun Bangay patiently provided
incomparable technical support in the installation and maintenance of my hardware and software,
and rescued me from more than one disaster when things went wrong. To Rhodes University I am
indebted for the use of computer facilities, and for granting me leave to complete the writing of the
book. And, of course, several generations of students have contributed in intangible ways by their
reaction to my courses.

The development of the software in this book relied heavily on the use of electronic mail, and I am
grateful to Randy Bush, compiler writer and network guru extraordinaire, for his friendship, and for
his help in making the Internet a reality in developing countries in Africa and elsewhere.
But, as always, the greatest debt is owed to my wife Sally and my children David and Helen, for
their love and support through the many hours when they must have wondered where my priorities
lay.

Pat Terry
Rhodes University
Grahamstown

Trademarks
Ada is a trademark of the US Department of Defense.
Apple II is a trademark of Apple Corporation.
Borland C++, Turbo C++, TurboPascal and Delphi are trademarks of Borland
International Corporation.
GNU C Compiler is a trademark of the Free Software Foundation.
IBM and IBM PC are trademarks of International Business Machines Corporation.
Intel is a registered trademark of Intel Corporation.
MC68000 and MC68020 are trademarks of Motorola Corporation.
MIPS is a trademark of MIPS computer systems.
Microsoft, MS and MS-DOS are registered trademarks and Windows is a trademark of
Microsoft Corporation.
SPARC is a trademark of Sun Microsystems.
Stony Brook Software and QuickMod are trademarks of Gogesch Micro Systems, Inc.
occam and Transputer are trademarks of Inmos.
UCSD Pascal and UCSD p-System are trademarks of the Regents of the University of
California.
UNIX is a registered trademark of AT&T Bell Laboratories.
Z80 is a trademark of Zilog Corporation.
 COMPILERS AND COMPILER
GENERATORS
an introduction with C++
© P.D. Terry, Rhodes University, 1996

e-mail p.terry@ru.ac.za

The Postscript ® edition of this book was derived from the on-line versions available at
http://www.scifac.ru.ac.za/compilers/, a WWW site that is occasionally updated, and which
contains the latest versions of the various editions of the book, with details of how to download
compressed versions of the text and its supporting software and courseware.

The original edition of this book, published originally by International Thomson, is now out of
print, but has a home page at http://cs.ru.ac.za/homes/cspt/compbook.htm. In preparing the on-line
edition, the opportunity was taken to correct the few typographical mistakes that crept into the first
printing, and to create a few hyperlinks to where the source files can be found.

Feel free to read and use this book for study or teaching, but please respect my copyright and do not
distribute it further without my consent. If you do make use of it I would appreciate hearing from
you.

CONTENTS
Preface

Acknowledgements

1 Introduction
1.1 Objectives
1.2 Systems programs and translators
1.3 The relationship between high-level languages and translators

2 Translator classification and structure

2.1 T-diagrams
2.2 Classes of translator
2.3 Phases in translation
2.4 Multi-stage translators
2.5 Interpreters, interpretive compilers, and emulators

3 Compiler construction and bootstrapping

3.1 Using a high-level host language
3.2 Porting a high-level translator
 3.3 Bootstrapping
3.4 Self-compiling compilers
3.5 The half bootstrap
3.6 Bootstrapping from a portable interpretive compiler
3.7 A P-code assembler

4 Machine emulation
4.1 Simple machine architecture
4.2 Addressing modes
4.3 Case study 1 - a single-accumulator machine
4.4 Case study 2 - a stack-oriented computer

5 Language specification
5.1 Syntax, semantics, and pragmatics
5.2 Languages, symbols, alphabets and strings
5.3 Regular expressions
5.4 Grammars and productions
5.5 Classic BNF notation for productions
5.6 Simple examples
5.7 Phrase structure and lexical structure
5.8 -productions
5.9 Extensions to BNF
5.10 Syntax diagrams
5.11 Formal treatment of semantics

6 Simple assemblers
6.1 A simple ASSEMBLER language
6.2 One- and two-pass assemblers, and symbol tables
6.3 Towards the construction of an assembler
6.4 Two-pass assembly
6.5 One-pass assembly

7 Advanced assembler features

7.1 Error detection
7.2 Simple expressions as addresses
7.3 Improved symbol table handling - hash tables
7.4 Macro-processing facilities
7.5 Conditional assembly
7.6 Relocatable code
7.7 Further projects

8 Grammars and their classification

8.1 Equivalent grammars
8.2 Case study - equivalent grammars for describing expressions
8.3 Some simple restrictions on grammars
 8.4 Ambiguous grammars
8.5 Context sensitivity
8.6 The Chomsky hierarchy
8.7 Case study - Clang

9 Deterministic top-down parsing

9.1 Deterministic top-down parsing
9.2 Restrictions on grammars so as to allow LL(1) parsing
9.3 The effect of the LL(1) conditions on language design

10 Parser and scanner construction

10.1 Construction of simple recursive descent parsers
10.2 Case studies
10.3 Syntax error detection and recovery
10.4 Construction of simple scanners
10.5 Case studies
10.6 LR parsing
10.7 Automated construction of scanners and parsers

11 Syntax-directed translation
11.1 Embedding semantic actions into syntax rules
11.2 Attribute grammars
11.3 Synthesized and inherited attributes
11.4 Classes of attribute grammars
11.5 Case study - a small student database

12 Using Coco/R - overview

12.1 Installing and running Coco/R
12.2 Case study - a simple adding machine
12.3 Scanner specification
12.4 Parser specification
12.5 The driver program

13 Using Coco/R - Case studies

13.1 Case study - Understanding C declarations
13.2 Case study - Generating one-address code from expressions
13.3 Case study - Generating one-address code from an AST
13.4 Case study - How do parser generators work?
13.5 Project suggestions

14 A simple compiler - the front end

14.1 Overall compiler structure
14.2 Source handling
14.3 Error reporting
 14.4 Lexical analysis
14.5 Syntax analysis
14.6 Error handling and constraint analysis
14.7 The symbol table handler
14.8 Other aspects of symbol table management - further types

15 A simple compiler - the back end

15.1 The code generation interface
15.2 Code generation for a simple stack machine
15.3 Other aspects of code generation

16 Simple block structure

16.1 Parameterless procedures
16.2 Storage management

17 Parameters and functions

17.1 Syntax and semantics
17.2 Symbol table support for context sensitive features
17.3 Actual parameters and stack frames
17.4 Hypothetical stack machine support for parameter passing
17.5 Context sensitivity and LL(1) conflict resolution
17.6 Semantic analysis and code generation
17.7 Language design issues

18 Concurrent programming
18.1 Fundamental concepts
18.2 Parallel processes, exclusion and synchronization
18.3 A semaphore-based system - syntax, semantics, and code generation
18.4 Run-time implementation

Appendix A: Software resources for this book

Appendix B: Source code for the Clang compiler/interpreter

Appendix C: Cocol grammar for the Clang compiler/interpreter

Appendix D: Source code for a macro assembler

Bibliography

Index
 Compilers and Compiler Generators © P.D. Terry, 2000

1 INTRODUCTION

1.1 Objectives
The use of computer languages is an essential link in the chain between human and computer. In
this text we hope to make the reader more aware of some aspects of

Imperative programming languages - their syntactic and semantic features; the ways of
specifying syntax and semantics; problem areas and ambiguities; the power and usefulness of
various features of a language.

Translators for programming languages - the various classes of translator (assemblers,

compilers, interpreters); implementation of translators.

Compiler generators - tools that are available to help automate the construction of translators
for programming languages.

This book is a complete revision of an earlier one published by Addison-Wesley (Terry, 1986). It
has been written so as not to be too theoretical, but to relate easily to languages which the reader
already knows or can readily understand, like Pascal, Modula-2, C or C++. The reader is expected
to have a good background in one of those languages, access to a good implementation of it, and,
preferably, some background in assembly language programming and simple machine architecture.
We shall rely quite heavily on this background, especially on the understanding the reader should
have of the meaning of various programming constructs.

Significant parts of the text concern themselves with case studies of actual translators for simple
languages. Other important parts of the text are to be found in the many exercises and suggestions
for further study and experimentation on the part of the reader. In short, the emphasis is on "doing"
rather than just "reading", and the reader who does not attempt the exercises will miss many, if not
most, of the finer points.

The primary language used in the implementation of our case studies is C++ (Stroustrup, 1990).
Machine readable source code for all these case studies is to be found on the IBM-PC compatible
diskette that is included with the book. As well as C++ versions of this code, we have provided
equivalent source in Modula-2 and Turbo Pascal, two other languages that are eminently suitable
for use in a course of this nature. Indeed, for clarity, some of the discussion is presented in a
pseudo-code that often resembles Modula-2 rather more than it does C++. It is only fair to warn the
reader that the code extracts in the book are often just that - extracts - and that there are many
instances where identifiers are used whose meaning may not be immediately apparent from their
local context. The conscientious reader will have to expend some effort in browsing the code.
Complete source for an assembler and interpreter appears in the appendices, but the discussion
often revolves around simplified versions of these programs that are found in their entirety only on
the diskette.
1.2 Systems programs and translators
Users of modern computing systems can be divided into two broad categories. There are those who
never develop their own programs, but simply use ones developed by others. Then there are those
who are concerned as much with the development of programs as with their subsequent use. This
latter group - of whom we as computer scientists form a part - is fortunate in that program
development is usually aided by the use of high-level languages for expressing algorithms, the use
of interactive editors for program entry and modification, and the use of sophisticated job control
languages or graphical user interfaces for control of execution. Programmers armed with such tools
have a very different picture of computer systems from those who are presented with the hardware
alone, since the use of compilers, editors and operating systems - a class of tools known generally
as systems programs - removes from humans the burden of developing their systems at the
machine level. That is not to claim that the use of such tools removes all burdens, or all possibilities
for error, as the reader will be well aware.

Well within living memory, much program development was done in machine language - indeed,
some of it, of necessity, still is - and perhaps some readers have even tried this for themselves when
experimenting with microprocessors. Just a brief exposure to programs written as almost
meaningless collections of binary or hexadecimal digits is usually enough to make one grateful for
the presence of high-level languages, clumsy and irritating though some of their features may be.

However, in order for high-level languages to be usable, one must be able to convert programs
written in them into the binary or hexadecimal digits and bitstrings that a machine will understand.
At an early stage it was realized that if constraints were put on the syntax of a high-level language
the translation process became one that could be automated. This led to the development of
translators or compilers - programs which accept (as data) a textual representation of an algorithm
expressed in a source language, and which produce (as primary output) a representation of the
same algorithm expressed in another language, the object or target language.

Beginners often fail to distinguish between the compilation (compile-time) and execution (run-time)
phases in developing and using programs written in high-level languages. This is an easy trap to fall
into, since the translation (compilation) is often hidden from sight, or invoked with a special
function key from within an integrated development environment that may possess many other
magic function keys. Furthermore, beginners are often taught programming with this distinction
deliberately blurred, their teachers offering explanations such as "when a computer executes a read
statement it reads a number from the input data into a variable". This hides several low-level
operations from the beginner. The underlying implications of file handling, character conversion,
and storage allocation are glibly ignored - as indeed is the necessity for the computer to be
programmed to understand the word read in the first place. Anyone who has attempted to program
input/output (I/O) operations directly in assembler languages will know that many of them are
non-trivial to implement.

A translator, being a program in its own right, must itself be written in a computer language, known
as its host or implementation language. Today it is rare to find translators that have been
developed from scratch in machine language. Clearly the first translators had to be written in this
way, and at the outset of translator development for any new system one has to come to terms with
the machine language and machine architecture for that system. Even so, translators for new
machines are now invariably developed in high-level languages, often using the techniques of
cross-compilation and bootstrapping that will be discussed in more detail later.

The first major translators written may well have been the Fortran compilers developed by Backus
and his colleagues at IBM in the 1950’s, although machine code development aids were in
existence by then. The first Fortran compiler is estimated to have taken about 18 person-years of
effort. It is interesting to note that one of the primary concerns of the team was to develop a system
that could produce object code whose efficiency of execution would compare favourably with that
which expert human machine coders could achieve. An automatic translation process can rarely
produce code as optimal as can be written by a really skilled user of machine language, and to this
day important components of systems are often developed at (or very near to) machine level, in the
interests of saving time or space.

Translator programs themselves are never completely portable (although parts of them may be), and
they usually depend to some extent on other systems programs that the user has at his or her
disposal. In particular, input/output and file management on modern computer systems are usually
controlled by the operating system. This is a program or suite of programs and routines whose job
it is to control the execution of other programs so as best to share resources such as printers,
plotters, disk files and tapes, often making use of sophisticated techniques such as parallel
processing, multiprogramming and so on. For many years the development of operating systems
required the use of programming languages that remained closer to the machine code level than did
languages suitable for scientific or commercial programming. More recently a number of successful
higher level languages have been developed with the express purpose of catering for the design of
operating systems and real-time control. The most obvious example of such a language is C,
developed originally for the implementation of the UNIX operating system, and now widely used in
all areas of computing.

1.3 The relationship between high-level languages and translators

The reader will rapidly become aware that the design and implementation of translators is a subject
that may be developed from many possible angles and approaches. The same is true for the design
of programming languages.

Computer languages are generally classed as being "high-level" (like Pascal, Fortran, Ada,
Modula-2, Oberon, C or C++) or "low-level" (like ASSEMBLER). High-level languages may
further be classified as "imperative" (like all of those just mentioned), or "functional" (like Lisp,
Scheme, ML, or Haskell), or "logic" (like Prolog).

High-level languages are claimed to possess several advantages over low-level ones:

Readability: A good high-level language will allow programs to be written that in some ways
resemble a quasi-English description of the underlying algorithms. If care is taken, the coding
may be done in a way that is essentially self-documenting, a highly desirable property when
one considers that many programs are written once, but possibly studied by humans many
times thereafter.

Portability: High-level languages, being essentially machine independent, hold out the
promise of being used to develop portable software. This is software that can, in principle
(and even occasionally in practice), run unchanged on a variety of different machines -
provided only that the source code is recompiled as it moves from machine to machine.

To achieve machine independence, high-level languages may deny access to low-level

features, and are sometimes spurned by programmers who have to develop low-level machine
dependent systems. However, some languages, like C and Modula-2, were specifically
designed to allow access to these features from within the context of high-level constructs.
 Structure and object orientation: There is general agreement that the structured programming
movement of the 1960’s and the object-oriented movement of the 1990’s have resulted in a
great improvement in the quality and reliability of code. High-level languages can be
designed so as to encourage or even subtly enforce these programming paradigms.

Generality: Most high-level languages allow the writing of a wide variety of programs, thus
relieving the programmer of the need to become expert in many diverse languages.

Brevity: Programs expressed in high-level languages are often considerably shorter (in terms
of their number of source lines) than their low-level equivalents.

Error checking: Being human, a programmer is likely to make many mistakes in the
development of a computer program. Many high-level languages - or at least their
implementations - can, and often do, enforce a great deal of error checking both at
compile-time and at run-time. For this they are, of course, often criticized by programmers
who have to develop time-critical code, or who want their programs to abort as quickly as
possible.

These advantages sometimes appear to be over-rated, or at any rate, hard to reconcile with reality.
For example, readability is usually within the confines of a rather stilted style, and some beginners
are disillusioned when they find just how unnatural a high-level language is. Similarly, the
generality of many languages is confined to relatively narrow areas, and programmers are often
dismayed when they find areas (like string handling in standard Pascal) which seem to be very
poorly handled. The explanation is often to be found in the close coupling between the development
of high-level languages and of their translators. When one examines successful languages, one finds
numerous examples of compromise, dictated largely by the need to accommodate language ideas to
rather uncompromising, if not unsuitable, machine architectures. To a lesser extent, compromise is
also dictated by the quirks of the interface to established operating systems on machines. Finally,
some appealing language features turn out to be either impossibly difficult to implement, or too
expensive to justify in terms of the machine resources needed. It may not immediately be apparent
that the design of Pascal (and of several of its successors such as Modula-2 and Oberon) was
governed partly by a desire to make it easy to compile. It is a tribute to its designer that, in spite of
the limitations which this desire naturally introduced, Pascal became so popular, the model for so
many other languages and extensions, and encouraged the development of superfast compilers such
as are found in Borland’s Turbo Pascal and Delphi systems.

The design of a programming language requires a high degree of skill and judgement. There is
evidence to show that one’s language is not only useful for expressing one’s ideas. Because
language is also used to formulate and develop ideas, one’s knowledge of language largely
determines how and, indeed, what one can think. In the case of programming languages, there has
been much controversy over this. For example, in languages like Fortran - for long the lingua
franca of the scientific computing community - recursive algorithms were "difficult" to use (not
impossible, just difficult!), with the result that many programmers brought up on Fortran found
recursion strange and difficult, even something to be avoided at all costs. It is true that recursive
algorithms are sometimes "inefficient", and that compilers for languages which allow recursion
may exacerbate this; on the other hand it is also true that some algorithms are more simply
explained in a recursive way than in one which depends on explicit repetition (the best examples
probably being those associated with tree manipulation).

There are two divergent schools of thought as to how programming languages should be designed.
The one, typified by the Wirth school, stresses that languages should be small and understandable,
and that much time should be spent in consideration of what tempting features might be omitted
without crippling the language as a vehicle for system development. The other, beloved of
languages designed by committees with the desire to please everyone, packs a language full of
every conceivable potentially useful feature. Both schools claim success. The Wirth school has
given us Pascal, Modula-2 and Oberon, all of which have had an enormous effect on the thinking of
computer scientists. The other approach has given us Ada, C and C++, which are far more difficult
to master well and extremely complicated to implement correctly, but which claim spectacular
successes in the marketplace.

Other aspects of language design that contribute to success include the following:

Orthogonality: Good languages tend to have a small number of well thought out features that
can be combined in a logical way to supply more powerful building blocks. Ideally these
features should not interfere with one another, and should not be hedged about by a host of
inconsistencies, exceptional cases and arbitrary restrictions. Most languages have blemishes -
for example, in Wirth’s original Pascal a function could only return a scalar value, not one of
any structured type. Many potentially attractive extensions to well-established languages
prove to be extremely vulnerable to unfortunate oversights in this regard.

Familiar notation: Most computers are "binary" in nature. Blessed with ten toes on which to
check out their number-crunching programs, humans may be somewhat relieved that
high-level languages usually make decimal arithmetic the rule, rather than the exception, and
provide for mathematical operations in a notation consistent with standard mathematics.
When new languages are proposed, these often take the form of derivatives or dialects of
well-established ones, so that programmers can be tempted to migrate to the new language
and still feel largely at home - this was the route taken in developing C++ from C, Java from
C++, and Oberon from Modula-2, for example.

Besides meeting the ones mentioned above, a successful modern high-level language will have
been designed to meet the following additional criteria:

Clearly defined: It must be clearly described, for the benefit of both the user and the compiler
writer.

Quickly translated: It should admit quick translation, so that program development time when
using the language is not excessive.

Modularity: It is desirable that programs can be developed in the language as a collection of

separately compiled modules, with appropriate mechanisms for ensuring self-consistency
between these modules.

Efficient: It should permit the generation of efficient object code.

Widely available: It should be possible to provide translators for all the major machines and
for all the major operating systems.

The importance of a clear language description or specification cannot be over-emphasized. This

must apply, firstly, to the so-called syntax of the language - that is, it must specify accurately what
form a source program may assume. It must apply, secondly, to the so-called static semantics of
the language - for example, it must be clear what constraints must be placed on the use of entities of
differing types, or the scope that various identifiers have across the program text. Finally, the
specification must also apply to the dynamic semantics of programs that satisfy the syntactic and
static semantic rules - that is, it must be capable of predicting the effect any program expressed in
that language will have when it is executed.

Programming language description is extremely difficult to do accurately, especially if it is

attempted through the medium of potentially confusing languages like English. There is an
increasing trend towards the use of formalism for this purpose, some of which will be illustrated in
later chapters. Formal methods have the advantage of precision, since they make use of the clearly
defined notations of mathematics. To offset this, they may be somewhat daunting to programmers
weak in mathematics, and do not necessarily have the advantage of being very concise - for
example, the informal description of Modula-2 (albeit slightly ambiguous in places) took only some
35 pages (Wirth, 1985), while a formal description prepared by an ISO committee runs to over 700
pages.

Formal specifications have the added advantage that, in principle, and to a growing degree in
practice, they may be used to help automate the implementation of translators for the language.
Indeed, it is increasingly rare to find modern compilers that have been implemented without the
help of so-called compiler generators. These are programs that take a formal description of the
syntax and semantics of a programming language as input, and produce major parts of a compiler
for that language as output. We shall illustrate the use of compiler generators at appropriate points
in our discussion, although we shall also show how compilers may be crafted by hand.

Exercises

1.1 Make a list of as many translators as you can think of that can be found on your computer
system.

1.2 Make a list of as many other systems programs (and their functions) as you can think of that can
be found on your computer system.

1.3 Make a list of existing features in your favourite (or least favourite) programming language that
you find irksome. Make a similar list of features that you would like to have seen added. Then
examine your lists and consider which of the features are probably related to the difficulty of
implementation.

Further reading

As we proceed, we hope to make the reader more aware of some of the points raised in this section.
Language design is a difficult area, and much has been, and continues to be, written on the topic.
The reader might like to refer to the books by Tremblay and Sorenson (1985), Watson (1989), and
Watt (1991) for readable summaries of the subject, and to the papers by Wirth (1974, 1976a,
1988a), Kernighan (1981), Welsh, Sneeringer and Hoare (1977), and Cailliau (1982). Interesting
background on several well-known languages can be found in ACM SIGPLAN Notices for August
1978 and March 1993 (Lee and Sammet, 1978, 1993), two special issues of that journal devoted to
the history of programming language development. Stroustrup (1993) gives a fascinating exposition
of the development of C++, arguably the most widely used language at the present time. The terms
"static semantics" and "dynamic semantics" are not used by all authors; for a discussion on this
point see the paper by Meek (1990).
 Compilers and Compiler Generators © P.D. Terry, 2000

2 TRANSLATOR CLASSIFICATION AND STRUCTURE

In this chapter we provide the reader with an overview of the inner structure of translators, and
some idea of how they are classified.

A translator may formally be defined as a function, whose domain is a source language, and whose
range is contained in an object or target language.

A little experience with translators will reveal that it is rarely considered part of the translator’s
function to execute the algorithm expressed by the source, merely to change its representation from
one form to another. In fact, at least three languages are involved in the development of translators:
the source language to be translated, the object or target language to be generated, and the host
language to be used for implementing the translator. If the translation takes place in several stages,
there may even be other, intermediate, languages. Most of these - and, indeed, the host language
and object languages themselves - usually remain hidden from a user of the source language.

2.1 T-diagrams
A useful notation for describing a computer program, particularly a translator, uses so-called
T-diagrams, examples of which are shown in Figure 2.1.

We shall use the notation "M-code" to stand for "machine code" in these diagrams. Translation
itself is represented by standing the T on a machine, and placing the source program and object
program on the left and right arms, as depicted in Figure 2.2.
We can also regard this particular combination as depicting an abstract machine (sometimes called
a virtual machine), whose aim in life is to convert Turbo Pascal source programs into their 8086
machine code equivalents.

T-diagrams were first introduced by Bratman (1961). They were further refined by Earley and
Sturgis (1970), and are also used in the books by Bennett (1990), Watt (1993), and Aho, Sethi and
Ullman (1986).

2.2 Classes of translator

It is common to distinguish between several well-established classes of translator:

The term assembler is usually associated with those translators that map low-level language
instructions into machine code which can then be executed directly. Individual source
language statements usually map one-for-one to machine-level instructions.

The term macro-assembler is also associated with those translators that map low-level
language instructions into machine code, and is a variation on the above. Most source
language statements map one- for-one into their target language equivalents, but some macro
statements map into a sequence of machine- level instructions - effectively providing a text
replacement facility, and thereby extending the assembly language to suit the user. (This is
not to be confused with the use of procedures or other subprograms to "extend" high-level
languages, because the method of implementation is usually very different.)

The term compiler is usually associated with those translators that map high-level language
instructions into machine code which can then be executed directly. Individual source
language statements usually map into many machine-level instructions.

The term pre-processor is usually associated with those translators that map a superset of a
high-level language into the original high-level language, or that perform simple text
substitutions before translation takes place. The best-known pre-processor is probably that
which forms an integral part of implementations of the language C, and which provides many
of the features that contribute to the widely- held perception that C is the only really portable
language.

The term high-level translator is often associated with those translators that map one
high-level language into another high-level language - usually one for which sophisticated
compilers already exist on a range of machines. Such translators are particularly useful as
components of a two-stage compiling system, or in assisting with the bootstrapping
techniques to be discussed shortly.
 The terms decompiler and disassembler refer to translators which attempt to take object
code at a low level and regenerate source code at a higher level. While this can be done quite
successfully for the production of assembler level code, it is much more difficult when one
tries to recreate source code originally written in, say, Pascal.

Many translators generate code for their host machines. These are called self-resident translators.
Others, known as cross-translators, generate code for machines other than the host machine.
Cross-translators are often used in connection with microcomputers, especially in embedded
systems, which may themselves be too small to allow self-resident translators to operate
satisfactorily. Of course, cross-translation introduces additional problems in connection with
transferring the object code from the donor machine to the machine that is to execute the translated
program, and can lead to delays and frustration in program development.

The output of some translators is absolute machine code, left loaded at fixed locations in a machine
ready for immediate execution. Other translators, known as load-and-go translators, may even
initiate execution of this code. However, a great many translators do not produce fixed-address
machine code. Rather, they produce something closely akin to it, known as semicompiled or
binary symbolic or relocatable form. A frequent use for this is in the development of composite
libraries of special purpose routines, possibly originating from a mixture of source languages.
Routines compiled in this way are linked together by programs called linkage editors or linkers,
which may be regarded almost as providing the final stage for a multi-stage translator. Languages
that encourage the separate compilation of parts of a program - like Modula-2 and C++ - depend
critically on the existence of such linkers, as the reader is doubtless aware. For developing really
large software projects such systems are invaluable, although for the sort of "throw away" programs
on which most students cut their teeth, they can initially appear to be a nuisance, because of the
overheads of managing several files, and of the time taken to link their contents together.

T-diagrams can be combined to show the interdependence of translators, loaders and so on. For
example, the FST Modula-2 system makes use of a compiler and linker as shown in Figure 2.3.

Exercises

2.1 Make a list of as many translators as you can think of that can be found on your system.

2.2 Which of the translators known to you are of the load-and-go type?

2.3 Do you know whether any of the translators you use produce relocatable code? Is this of a
standard form? Do you know the names of the linkage editors or loaders used on your system?
2.4 Are there any pre-processors on your system? What are they used for?

2.3 Phases in translation

Translators are highly complex programs, and it is unreasonable to consider the translation process
as occurring in a single step. It is usual to regard it as divided into a series of phases. The simplest
breakdown recognizes that there is an analytic phase, in which the source program is analysed to
determine whether it meets the syntactic and static semantic constraints imposed by the language.
This is followed by a synthetic phase in which the corresponding object code is generated in the
target language. The components of the translator that handle these two major phases are said to
comprise the front end and the back end of the compiler. The front end is largely independent of
the target machine, the back end depends very heavily on the target machine. Within this structure
we can recognize smaller components or phases, as shown in Figure 2.4.

The character handler is the section that communicates with the outside world, through the
operating system, to read in the characters that make up the source text. As character sets and file
handling vary from system to system, this phase is often machine or operating system dependent.

The lexical analyser or scanner is the section that fuses characters of the source text into groups
that logically make up the tokens of the language - symbols like identifiers, strings, numeric
constants, keywords like while and if, operators like <=, and so on. Some of these symbols are
very simply represented on the output from the scanner, some need to be associated with various
properties such as their names or values.

Lexical analysis is sometimes easy, and at other times not. For example, the Modula-2 statement
WHILE A > 3 * B DO A := A - 1 END

easily decodes into tokens

WHILE keyword
A identifier name A
 > operator comparison
3 constant literal value 3
* operator multiplication
B identifier name B
DO keyword
A identifier name A
:= operator assignment
A identifier name A
- operator subtraction
1 constant literal value 1
END keyword

as we read it from left to right, but the Fortran statement

10 DO 20 I = 1 . 30

is more deceptive. Readers familiar with Fortran might see it as decoding into
10 label
DO keyword
20 statement label
I INTEGER identifier
= assignment operator
1 INTEGER constant literal
, separator
30 INTEGER constant literal

while those who enjoy perversity might like to see it as it really is:
10 label
DO20I REAL identifier
= assignment operator
1.30 REAL constant literal

One has to look quite hard to distinguish the period from the "expected" comma. (Spaces are
irrelevant in Fortran; one would, of course be perverse to use identifiers with unnecessary and
highly suggestive spaces in them.) While languages like Pascal, Modula-2 and C++ have been
cleverly designed so that lexical analysis can be clearly separated from the rest of the analysis, the
same is obviously not true of Fortran and other languages that do not have reserved keywords.

The syntax analyser or parser groups the tokens produced by the scanner into syntactic structures
- which it does by parsing expressions and statements. (This is analogous to a human analysing a
sentence to find components like "subject", "object" and "dependent clauses"). Often the parser is
combined with the contextual constraint analyser, whose job it is to determine that the
components of the syntactic structures satisfy such things as scope rules and type rules within the
context of the structure being analysed. For example, in Modula-2 the syntax of a while statement is
sometimes described as
WHILE Expression DO StatementSequence END

It is reasonable to think of a statement in the above form with any type of Expression as being
syntactically correct, but as being devoid of real meaning unless the value of the Expression is
constrained (in this context) to be of the Boolean type. No program really has any meaning until it
is executed dynamically. However, it is possible with strongly typed languages to predict at
compile-time that some source programs can have no sensible meaning (that is, statically, before an
attempt is made to execute the program dynamically). Semantics is a term used to describe
"meaning", and so the constraint analyser is often called the static semantic analyser, or simply
the semantic analyser.

The output of the syntax analyser and semantic analyser phases is sometimes expressed in the form
of a decorated abstract syntax tree (AST). This is a very useful representation, as it can be used in
clever ways to optimize code generation at a later stage.
Whereas the concrete syntax of many programming languages incorporates many keywords and
tokens, the abstract syntax is rather simpler, retaining only those components of the language
needed to capture the real content and (ultimately) meaning of the program. For example, whereas
the concrete syntax of a while statement requires the presence of WHILE, DO and END as shown
above, the essential components of the while statement are simply the (Boolean) Expression and the
statements comprising the StatementSequence.

Thus the Modula-2 statement

WHILE (1 < P) AND (P < 9) DO P := P + Q END

or its C++ equivalent

while (1 < P && P < 9) P = P + Q;

are both depicted by the common AST shown in Figure 2.5.

An abstract syntax tree on its own is devoid of some semantic detail; the semantic analyser has the
task of adding "type" and other contextual information to the various nodes (hence the term
"decorated" tree).

Sometimes, as for example in the case of most Pascal compilers, the construction of such a tree is
not explicit, but remains implicit in the recursive calls to procedures that perform the syntax and
semantic analysis.

Of course, it is also possible to construct concrete syntax trees. The Modula-2 form of the statement
WHILE (1 < P) AND (P < 9) DO P := P + Q END

could be depicted in full and tedious detail by the tree shown in Figure 2.6. The reader may have to
make reference to Modula-2 syntax diagrams and the knowledge of Modula-2 precedence rules to
understand why the tree looks so complicated.
The phases just discussed are all analytic in nature. The ones that follow are more synthetic. The
first of these might be an intermediate code generator, which, in practice, may also be integrated
with earlier phases, or omitted altogether in the case of some very simple translators. It uses the
data structures produced by the earlier phases to generate a form of code, perhaps in the form of
simple code skeletons or macros, or ASSEMBLER or even high-level code for processing by an
external assembler or separate compiler. The major difference between intermediate code and
actual machine code is that intermediate code need not specify in detail such things as the exact
machine registers to be used, the exact addresses to be referred to, and so on.

Our example statement

WHILE (1 < P) AND (P < 9) DO P := P + Q END

might produce intermediate code equivalent to

L0 if 1 < P goto L1
goto L3
L1 if P < 9 goto L2
goto L3
L2 P := P + Q
goto L0
L3 continue

Then again, it might produce something like

L0 T1 := 1 < P
T2 := P < 9
if T1 and T2 goto L1
goto L2
L1 P := P + Q
goto L0
L2 continue

depending on whether the implementors of the translator use the so-called sequential conjunction or
short-circuit approach to handling compound Boolean expressions (as in the first case) or the
so-called Boolean operator approach. The reader will recall that Modula-2 and C++ require the
short-circuit approach. However, the very similar language Pascal did not specify that one approach
be preferred above the other.

A code optimizer may optionally be provided, in an attempt to improve the intermediate code in
the interests of speed or space or both. To use the same example as before, obvious optimization
would lead to code equivalent to
L0 if 1 >= P goto L1
if P >= 9 goto L1
P := P + Q
goto L0
L1 continue

The most important phase in the back end is the responsibility of the code generator. In a real
compiler this phase takes the output from the previous phase and produces the object code, by
deciding on the memory locations for data, generating code to access such locations, selecting
registers for intermediate calculations and indexing, and so on. Clearly this is a phase which calls
for much skill and attention to detail, if the finished product is to be at all efficient. Some translators
go on to a further phase by incorporating a so-called peephole optimizer in which attempts are
made to reduce unnecessary operations still further by examining short sequences of generated code
in closer detail.

Below we list the actual code generated by various MS-DOS compilers for this statement. It is
readily apparent that the code generation phases in these compilers are markedly different. Such
differences can have a profound effect on program size and execution speed.
Borland C++ 3.1 (47 bytes) Turbo Pascal (46 bytes)
(with no short circuit evaluation)

CS:A0 BBB702 MOV BX,02B7 CS:09 833E3E0009 CMP WORD PTR[003E],9

CS:A3 C746FE5100 MOV WORD PTR[BP-2],0051 CS:0E 7C04 JL 14
CS:A8 EB07 JMP B1 CS:10 B000 MOV AL,0
CS:AA 8BC3 MOV AX,BX CS:12 EB02 JMP 16
CS:AC 0346FE ADD AX,[BP-2] CS:14 B001 MOV AL,1
CS:AF 8BD8 MOV BX,AX CS:16 8AD0 MOV DL,AL
CS:B1 83FB01 CMP BX,1 CS:18 833E3E0001 CMP WORD PTR[003E],1
CS:B4 7E05 JLE BB CS:1D 7F04 JG 23
CS:B6 B80100 MOV AX,1 CS:1F B000 MOV AL,0
CS:B9 EB02 JMP BD CS:21 EB02 JMP 25
CS:BB 33C0 XOR AX,AX CS:23 B001 MOV AL,01
CS:BD 50 PUSH AX CS:25 22C2 AND AL,DL
CS:BE 83FB09 CMP BX,9 CS:27 08C0 OR AL,AL
CS:C1 7D05 JGE C8 CS:29 740C JZ 37
CS:C3 B80100 MOV AX,1 CS:2B A13E00 MOV AX,[003E]
CS:C6 EB02 JMP CA CS:2E 03064000 ADD AX,[0040]
CS:C8 33C0 XOR AX,AX CS:32 A33E00 MOV [003E],AX
CS:CA 5A POP DX CS:35 EBD2 JMP 9
CS:CB 85D0 TEST DX,AX
CS:CD 75DB JNZ AA

JPI TopSpeed Modula-2 (29 bytes) Stony Brook QuickMod (24 bytes)

CS:19 2E CS: CS:69 BB2D00 MOV BX,2D

CS:1A 8E1E2700 MOV DS,[0027] CS:6C B90200 MOV CX,2
CS:1E 833E000001 CMP WORD PTR[0000],1 CS:6F E90200 JMP 74
CS:23 7E11 JLE 36 CS:72 01D9 ADD CX,BX
CS:25 833E000009 CMP WORD PTR[0000],9 CS:74 83F901 CMP CX,1
CS:2A 7D0A JGE 36 CS:77 7F03 JG 7C
CS:2C 8B0E0200 MOV CX,[0002] CS:79 E90500 JMP 81
CS:30 010E0000 ADD [0000],CX CS:7C 83F909 CMP CX,9
CS:34 EBE3 JMP 19 CS:7F 7CF1 JL 72

A translator inevitably makes use of a complex data structure, known as the symbol table, in which
it keeps track of the names used by the program, and associated properties for these, such as their
type, and their storage requirements (in the case of variables), or their values (in the case of
constants).
As is well known, users of high-level languages are apt to make many errors in the development of
even quite simple programs. Thus the various phases of a compiler, especially the earlier ones, also
communicate with an error handler and error reporter which are invoked when errors are
detected. It is desirable that compilation of erroneous programs be continued, if possible, so that the
user can clean several errors out of the source before recompiling. This raises very interesting
issues regarding the design of error recovery and error correction techniques. (We speak of error
recovery when the translation process attempts to carry on after detecting an error, and of error
correction or error repair when it attempts to correct the error from context - usually a contentious
subject, as the correction may be nothing like what the programmer originally had in mind.)

Error detection at compile-time in the source code must not be confused with error detection at
run-time when executing the object code. Many code generators are responsible for adding
error-checking code to the object program (to check that subscripts for arrays stay in bounds, for
example). This may be quite rudimentary, or it may involve adding considerable code and data
structures for use with sophisticated debugging systems. Such ancillary code can drastically reduce
the efficiency of a program, and some compilers allow it to be suppressed.

Sometimes mistakes in a program that are detected at compile-time are known as errors, and errors
that show up at run-time are known as exceptions, but there is no universally agreed terminology
for this.

Figure 2.4 seems to imply that compilers work serially, and that each phase communicates with the
next by means of a suitable intermediate language, but in practice the distinction between the
various phases often becomes a little blurred. Moreover, many compilers are actually constructed
around a central parser as the dominant component, with a structure rather more like the one in
Figure 2.7.

Exercises

2.5 What sort of problems can you foresee a Fortran compiler having in analysing statements
beginning
IF ( I(J) - I(K) ) ........
CALL IF (4 , ...........
IF (3 .EQ. MAX) GOTO ......
100 FORMAT(X3H)=(I5)

2.6 What sort of code would you have produced had you been coding a statement like "WHILE (1 <
P) AND (P < 9) DO P := P + Q END" into your favourite ASSEMBLER language?

2.7 Draw the concrete syntax tree for the C++ version of the while statement used for illustration in
this section.

2.8 Are there any reasons why short-circuit evaluation should be preferred over the Boolean
operator approach? Can you think of any algorithms that would depend critically on which
approach was adopted?

2.9 Write down a few other high-level constructs and try to imagine what sort of
ASSEMBLER-like machine code a compiler would produce for them.

2.10 What do you suppose makes it relatively easy to compile Pascal? Can you think of any aspects
of Pascal which could prove really difficult?

2.11 We have used two undefined terms which at first seem interchangeable, namely "separate" and
"independent" compilation. See if you can discover what the differences are.

2.12 Many development systems - in particular debuggers - allow a user to examine the object code
produced by a compiler. If you have access to one of these, try writing a few very simple (single
statement) programs, and look at the sort of object code that is generated for them.

2.4 Multi-stage translators

Besides being conceptually divided into phases, translators are often divided into passes, in each of
which several phases may be combined or interleaved. Traditionally, a pass reads the source
program, or output from a previous pass, makes some transformations, and then writes output to an
intermediate file, whence it may be rescanned on a subsequent pass.

These passes may be handled by different integrated parts of a single compiler, or they may be
handled by running two or more separate programs. They may communicate by using their own
specialized forms of intermediate language, they may communicate by making use of internal data
structures (rather than files), or they may make several passes over the same original source code.

The number of passes used depends on a variety of factors. Certain languages require at least two
passes to be made if code is to be generated easily - for example, those where declaration of
identifiers may occur after the first reference to the identifier, or where properties associated with
an identifier cannot be readily deduced from the context in which it first appears. A multi-pass
compiler can often save space. Although modern computers are usually blessed with far more
memory than their predecessors of only a few years back, multiple passes may be an important
consideration if one wishes to translate complicated languages within the confines of small systems.
Multi-pass compilers may also allow for better provision of code optimization, error reporting and
error handling. Lastly, they lend themselves to team development, with different members of the
team assuming responsibility for different passes. However, multi-pass compilers are usually
slower than single-pass ones, and their probable need to keep track of several files makes them
slightly awkward to write and to use. Compromises at the design stage often result in languages that
are well suited to single-pass compilation.

In practice, considerable use is made of two-stage translators in which the first stage is a high-level
translator that converts the source program into ASSEMBLER, or even into some other relatively
high-level language for which an efficient translator already exists. The compilation process would
then be depicted as in Figure 2.8 - our example shows a Modula-3 program being prepared for
execution on a machine that has a Modula-3 to C converter:

It is increasingly common to find compilers for high-level languages that have been implemented
using C, and which themselves produce C code as output. The success of these is based on the
premises that "all modern computers come equipped with a C compiler" and "source code written in
C is truly portable". Neither premise is, unfortunately, completely true. However, compilers written
in this way are as close to achieving the dream of themselves being portable as any that exist at the
present time. The way in which such compilers may be used is discussed further in Chapter 3.

Exercises

2.13 Try to find out which of the compilers you have used are single-pass, and which are
multi-pass, and for the latter, find out how many passes are involved. Which produce relocatable
code needing further processing by linkers or linkage editors?

2.14 Do any of the compilers in use on your system produce ASSEMBLER, C or other such code
during the compilation process? Can you foresee any particular problems that users might
experience in using such compilers?

2.15 One of several compilers that translates from Modula-2 to C is called mtc, and is freely
available from several ftp sites. If you are a Modula-2 programmer, obtain a copy, and experiment
with it.

2.16 An excellent compiler that translates Pascal to C is called p2c, and is widely available for Unix
systems from several ftp sites. If you are a Pascal programmer, obtain a copy, and experiment with
it.

2.17 Can you foresee any practical difficulties in using C as an intermediate language?

2.5 Interpreters, interpretive compilers, and emulators

Compilers of the sort that we have been discussing have a few properties that may not immediately
be apparent. Firstly, they usually aim to produce object code that can run at the full speed of the
target machine. Secondly, they are usually arranged to compile an entire section of code before any
of it can be executed.
In some interactive environments the need arises for systems that can execute part of an application
without preparing all of it, or ones that allow the user to vary his or her course of action on the fly.
Typical scenarios involve the use of spreadsheets, on-line databases, or batch files or shell scripts
for operating systems. With such systems it may be feasible (or even desirable) to exchange some
of the advantages of speed of execution for the advantage of procuring results on demand.

Systems like these are often constructed so as to make use of an interpreter. An interpreter is a
translator that effectively accepts a source program and executes it directly, without, seemingly,
producing any object code first. It does this by fetching the source program instructions one by one,
analysing them one by one, and then "executing" them one by one. Clearly, a scheme like this, if it
is to be successful, places some quite severe constraints on the nature of the source program.
Complex program structures such as nested procedures or compound statements do not lend
themselves easily to such treatment. On the other hand, one-line queries made of a data base, or
simple manipulations of a row or column of a spreadsheet, can be handled very effectively.

This idea is taken quite a lot further in the development of some translators for high-level
languages, known as interpretive compilers. Such translators produce (as output) intermediate
code which is intrinsically simple enough to satisfy the constraints imposed by a practical
interpreter, even though it may still be quite a long way from the machine code of the system on
which it is desired to execute the original program. Rather than continue translation to the level of
machine code, an alternative approach that may perform acceptably well is to use the intermediate
code as part of the input to a specially written interpreter. This in turn "executes" the original
algorithm, by simulating a virtual machine for which the intermediate code effectively is the
machine code. The distinction between the machine code and pseudo-code approaches to execution
is summarized in Figure 2.9.

We may depict the process used in an interpretive compiler running under MS-DOS for a toy
language like Clang, the one illustrated in later chapters, in T-diagram form (see Figure 2.10).

It is not necessary to confine interpreters merely to work with intermediate output from a translator.
More generally, of course, even a real machine can be viewed as a highly specialized interpreter -
one that executes the machine level instructions by fetching, analysing, and then interpreting them
one by one. In a real machine this all happens "in hardware", and hence very quickly. By carrying
on this train of thought, the reader should be able to see that a program could be written to allow
one real machine to emulate any other real machine, albeit perhaps slowly, simply by writing an
interpreter - or, as it is more usually called, an emulator - for the second machine.
For example, we might develop an emulator that runs on a Sun SPARC machine and makes it
appear to be an IBM PC (or the other way around). Once we have done this, we are (in principle) in
a position to execute any software developed for an IBM PC on the Sun SPARC machine -
effectively the PC software becomes portable!

The T-diagram notation is easily extended to handle the concept of such virtual machines. For
example, running Turbo Pascal on our Sun SPARC machine could be depicted by Figure 2.11.

The interpreter/emulator approach is widely used in the design and development both of new
machines themselves, and the software that is to run on those machines.

An interpretive approach may have several points in its favour:

It is far easier to generate hypothetical machine code (which can be tailored towards the
quirks of the original source language) than real machine code (which has to deal with the
uncompromising quirks of real machines).

A compiler written to produce (as output) well-defined pseudo-machine code capable of easy
interpretation on a range of machines can be made highly portable, especially if it is written in
a host language that is widely available (such as ANSI C), or even if it is made available
already implemented in its own pseudo- code.

It can more easily be made "user friendly" than can the native code approach. Since the
interpreter works closer to the source code than does a fully translated program, error
messages and other debugging aids may readily be related to this source.

A whole range of languages may quickly be implemented in a useful form on a wide range of
different machines relatively easily. This is done by producing intermediate code to a
well-defined standard, for which a relatively efficient interpreter should be easy to implement
on any particular real machine.

It proves to be useful in connection with cross-translators such as were mentioned earlier. The
code produced by such translators can sometimes be tested more effectively by simulated
execution on the donor machine, rather than after transfer to the target machine - the delays
inherent in the transfer from one machine to the other may be balanced by the degradation of
execution time in an interpretive simulation.

Lastly, intermediate languages are often very compact, allowing large programs to be
handled, even on relatively small machines. The success of the once very widely used UCSD
Pascal and UCSD p-System stands as an example of what can be done in this respect.
For all these advantages, interpretive systems carry fairly obvious overheads in execution speed,
because execution of intermediate code effectively carries with it the cost of virtual translation into
machine code each time a hypothetical machine instruction is obeyed.

One of the best known of the early portable interpretive compilers was the one developed at
Zürich and known as the "Pascal-P" compiler (Nori et al., 1981). This was supplied in a kit of three
components:

The first component was the source form of a Pascal compiler, written in a very complete
subset of the language, known as Pascal-P. The aim of this compiler was to translate Pascal-P
source programs into a well-defined and well-documented intermediate language, known as
P-code, which was the "machine code" for a hypothetical stack-based computer, known as the
P-machine.

The second component was a compiled version of the first - the P-codes that would be
produced by the Pascal-P compiler, were it to compile itself.

Lastly, the kit contained an interpreter for the P-code language, supplied as a Pascal
algorithm.

The interpreter served primarily as a model for writing a similar program for the target machine, to
allow it to emulate the hypothetical P-machine. As we shall see in a later chapter, emulators are
relatively easy to develop - even, if necessary, in ASSEMBLER - so that this stage was usually
fairly painlessly achieved. Once one had loaded the interpreter - that is to say, the version of it
tailored to a local real machine - into a real machine, one was in a position to "execute" P-code, and
in particular the P-code of the P-compiler. The compilation and execution of a user program could
then be achieved in a manner depicted in Figure 2.12.

Exercises

2.18 Try to find out which of the translators you have used are interpreters, rather than full
compilers.

2.19 If you have access to both a native-code compiler and an interpreter for a programming
language known to you, attempt to measure the loss in efficiency when the interpreter is used to run
a large program (perhaps one that does substantial number-crunching).
 Compilers and Compiler Generators © P.D. Terry, 2000

3 COMPILER CONSTRUCTION AND BOOTSTRAPPING

By now the reader may have realized that developing translators is a decidedly non-trivial exercise.
If one is faced with the task of writing a full-blown translator for a fairly complex source language,
or an emulator for a new virtual machine, or an interpreter for a low-level intermediate language,
one would probably prefer not to implement it all in machine code.

Fortunately one rarely has to contemplate such a radical step. Translator systems are now widely
available and well understood. A fairly obvious strategy when a translator is required for an old
language on a new machine, or a new language on an old machine (or even a new language on a
new machine), is to make use of existing compilers on either machine, and to do the development in
a high level language. This chapter provides a few examples that should make this clearer.

3.1 Using a high-level host language

If, as is increasingly common, one’s dream machine M is supplied with the machine coded version
of a compiler for a well-established language like C, then the production of a compiler for one’s
dream language X is achievable by writing the new compiler, say XtoM, in C and compiling the
source (XtoM.C) with the C compiler (CtoM.M) running directly on M (see Figure 3.1). This
produces the object version (XtoM.M) which can then be executed on M.

Even though development in C is much easier than development in machine code, the process is
still complex. As was mentioned earlier, it may be possible to develop a large part of the compiler
source using compiler generator tools - assuming, of course, that these are already available either
in executable form, or as C source that can itself be compiled easily. The hardest part of the
development is probably that associated with the back end, since this is intensely machine
dependent. If one has access to the source code of a compiler like CtoM one may be able to use this
to good avail. Although commercial compilers are rarely released in source form, source code is
available for many compilers produced at academic institutions or as components of the GNU
project carried out under the auspices of the Free Software Foundation.

3.2 Porting a high level translator

The process of modifying an existing compiler to work on a new machine is often known as
porting the compiler. In some cases this process may be almost trivially easy. Consider, for
example, the fairly common scenario where a compiler XtoC for a popular language X has been
implemented in C on machine A by writing a high-level translator to convert programs written in X
to C, and where it is desired to use language X on a machine M that, like A, has already been
blessed with a C compiler of its own. To construct a two-stage compiler for use on either machine,
all one needs to do, in principle, is to install the source code for XtoC on machine M and recompile
it.

Such an operation is conveniently represented in terms of T-diagrams chained together. Figure

3.2(a) shows the compilation of the X to C compiler, and Figure 3.2(b) shows the two-stage
compilation process needed to compile programs written in X to M-code.

The portability of a compiler like XtoC.C is almost guaranteed, provided that it is itself written in
"portable" C. Unfortunately, or as Mr. Murphy would put it, "interchangeable parts don’t" (more
explicitly, "portable C isn’t"). Some time may have to be spent in modifying the source code of
XtoC.C before it is acceptable as input to CtoM.M, although it is to be hoped that the developers of
XtoC.C will have used only standard C in their work, and used pre-processor directives that allow
for easy adaptation to other systems.

If there is an initial strong motivation for making a compiler portable to other systems it is, indeed,
often written so as to produce high-level code as output. More often, of course, the original
implementation of a language is written as a self-resident translator with the aim of directly
producing machine code for the current host system.

3.3 Bootstrapping
All this may seem to be skirting around a really nasty issue - how might the first high-level
language have been implemented? In ASSEMBLER? But then how was the assembler for
ASSEMBLER produced?

A full assembler is itself a major piece of software, albeit rather simple when compared with a
compiler for a really high level language, as we shall see. It is, however, quite common to define
one language as a subset of another, so that subset 1 is contained in subset 2 which in turn is
contained in subset 3 and so on, that is:
One might first write an assembler for subset 1 of ASSEMBLER in machine code, perhaps on a
load-and-go basis (more likely one writes in ASSEMBLER, and then hand translates it into
machine code). This subset assembler program might, perhaps, do very little other than convert
mnemonic opcodes into binary form. One might then write an assembler for subset 2 of
ASSEMBLER in subset 1 of ASSEMBLER, and so on.

This process, by which a simple language is used to translate a more complicated program, which
in turn may handle an even more complicated program and so on, is known as bootstrapping, by
analogy with the idea that it might be possible to lift oneself off the ground by tugging at one’s
boot-straps.

3.4 Self-compiling compilers

Once one has a working system, one can start using it to improve itself. Many compilers for popular
languages were first written in another implementation language, as implied in section 3.1, and then
rewritten in their own source language. The rewrite gives source for a compiler that can then be
compiled with the compiler written in the original implementation language. This is illustrated in
Figure 3.3.

Clearly, writing a compiler by hand not once, but twice, is a non-trivial operation, unless the
original implementation language is close to the source language. This is not uncommon: Oberon
compilers could be implemented in Modula-2; Modula-2 compilers, in turn, were first implemented
in Pascal (all three are fairly similar), and C++ compilers were first implemented in C.

Developing a self-compiling compiler has four distinct points to recommend it. Firstly, it
constitutes a non-trivial test of the viability of the language being compiled. Secondly, once it has
been done, further development can be done without recourse to other translator systems. Thirdly,
any improvements that can be made to its back end manifest themselves both as improvements to
the object code it produces for general programs and as improvements to the compiler itself. Lastly,
it provides a fairly exhaustive self-consistency check, for if the compiler is used to compile its own
source code, it should, of course, be able to reproduce its own object code (see Figure 3.4).

Furthermore, given a working compiler for a high-level language it is then very easy to produce
compilers for specialized dialects of that language.
3.5 The half bootstrap
Compilers written to produce object code for a particular machine are not intrinsically portable.
However, they are often used to assist in a porting operation. For example, by the time that the first
Pascal compiler was required for ICL machines, the Pascal compiler available in Zürich (where
Pascal had first been implemented on CDC mainframes) existed in two forms (Figure 3.5).

The first stage of the transportation process involved changing PasToCDC.Pas to generate ICL
machine code - thus producing a cross compiler. Since PasToCDC.Pas had been written in a
high-level language, this was not too difficult to do, and resulted in the compiler PasToICL.Pas.

Of course this compiler could not yet run on any machine at all. It was first compiled using
PasToCDC.CDC, on the CDC machine (see Figure 3.6(a)). This gave a cross-compiler that could
run on CDC machines, but still not, of course, on ICL machines. One further compilation of
PasToICL.Pas, using the cross-compiler PasToICL.CDC on the CDC machine, produced the final
result, PasToICL.ICL (Figure 3.6(b)).
The final product (PasToICL.ICL) was then transported on magnetic tape to the ICL machine, and
loaded quite easily. Having obtained a working system, the ICL team could (and did) continue
development of the system in Pascal itself.

This porting operation was an example of what is known as a half bootstrap system. The work of
transportation is essentially done entirely on the donor machine, without the need for any translator
in the target machine, but a crucial part of the original compiler (the back end, or code generator)
has to be rewritten in the process. Clearly the method is hazardous - any flaws or oversights in
writing PasToICL.Pas could have spelled disaster. Such problems can be reduced by minimizing
changes made to the original compiler. Another technique is to write an emulator for the target
machine that runs on the donor machine, so that the final compiler can be tested on the donor
machine before being transferred to the target machine.

3.6 Bootstrapping from a portable interpretive compiler

Because of the inherent difficulty of the half bootstrap for porting compilers, a variation on the full
bootstrap method described above for assemblers has often been successfully used in the case of
Pascal and other similar high-level languages. Here most of the development takes place on the
target machine, after a lot of preliminary work has been done on the donor machine to produce an
interpretive compiler that is almost portable. It will be helpful to illustrate with the well-known
example of the Pascal-P implementation kit mentioned in section 2.5.

Users of this kit typically commenced operations by implementing an interpreter for the P-machine.
The bootstrap process was then initiated by developing a compiler (PasPtoM.PasP) to translate
Pascal-P source programs to the local machine code. This compiler could be written in Pascal-P
source, development being guided by the source of the Pascal-P to P-code compiler supplied as part
of the kit. This new compiler was then compiled with the interpretive compiler (PasPtoP.P) from
the kit (Figure 3.7(a)) and the source of the Pascal to M-code compiler was then compiled by this
new compiler, interpreted once again by the P-machine, to give the final product, PasPtoM.M
(Figure 3.7(b)).

The Zürich P-code interpretive compiler could be, and indeed was, used as a highly portable
development system. It was employed to remarkable effect in developing the UCSD Pascal system,
which was the first serious attempt to implement Pascal on microcomputers. The UCSD Pascal
team went on to provide the framework for an entire operating system, editors and other utilities -
all written in Pascal, and all compiled into a well-defined P-code object code. Simply by providing
an alternative interpreter one could move the whole system to a new microcomputer system
virtually unchanged.

3.7 A P-code assembler

There is, of course, yet another way in which a portable interpretive compiler kit might be used.
One might commence by writing a P-code to M-code assembler, probably a relatively simple task.
Once this has been produced one would have the assembler depicted in Figure 3.8.

The P-codes for the P-code compiler would then be assembled by this system to give another cross
compiler (Figure 3.9(a)), and the same P-code/M-code assembler could then be used as a back-end
to the cross compiler (Figure 3.9(b)).

Exercises

3.1 Draw the T-diagram representations for the development of a P-code to M-code assembler,
assuming that you have a C++ compiler available on the target system.

3.2 Later in this text we shall develop an interpretive compiler for a small language called Clang,
using C++ as the host language. Draw T-diagram representations of the various components of the
system as you foresee them.

Further reading

A very clear exposition of bootstrapping is to be found in the book by Watt (1993). The ICL
bootstrap is further described by Welsh and Quinn (1972). Other early insights into bootstrapping
are to be found in papers by Lecarme and Peyrolle-Thomas (1973), by Nori et al. (1981), and
Cornelius, Lowman and Robson (1984).
 Compilers and Compiler Generators © P.D. Terry, 2000

4 MACHINE EMULATION
In Chapter 2 we discussed the use of emulation or interpretation as a tool for programming
language translation. In this chapter we aim to discuss hypothetical machine languages and the
emulation of hypothetical machines for these languages in more detail. Modern computers are
among the most complex machines ever designed by the human mind. However, this is a text on
programming language translation and not on electronic engineering, and our restricted discussion
will focus only on rather primitive object languages suited to the simple translators to be discussed
in later chapters.

4.1 Simple machine architecture

Many CPU (central processor unit) chips used in modern computers have one or more internal
registers or accumulators, which may be regarded as highly local memory where simple
arithmetic and logical operations may be performed, and between which local data transfers may
take place. These registers may be restricted to the capacity of a single byte (8 bits), or, as is typical
of most modern processors, they may come in a variety of small multiples of bytes or machine
words.

One fundamental internal register is the instruction register (IR), through which moves the
bitstrings (bytes) representing the fundamental machine-level instructions that the processor can
obey. These instructions tend to be extremely simple - operations such as "clear a register" or
"move a byte from one register to another" being the typical order of complexity. Some of these
instructions may be completely defined by a single byte value. Others may need two or more bytes
for a complete definition. Of these multi-byte instructions, the first usually denotes an operation,
and the rest relate either to a value to be operated upon, or to the address of a location in memory at
which can be found the value to be operated upon.

The simplest processors have only a few data registers, and are very limited in what they can
actually do with their contents, and so processors invariably make provision for interfacing to the
memory of the computer, and allow transfers to take place along so-called bus lines between the
internal registers and the far greater number of external memory locations. When information is to
be transferred to or from memory, the CPU places the appropriate address information on the
address bus, and then transmits or receives the data itself on the data bus. This is illustrated in
Figure 4.1.
The memory may simplistically be viewed as a one-dimensional array of byte values, analogous to
what might be described in high-level language terms by declarations like the following
TYPE
ADDRESS = CARDINAL [0 .. MemSize - 1];
BYTES = CARDINAL [0 .. 255];
VAR
Mem : ARRAY ADDRESS OF BYTES;

in Modula-2, or, in C++ (which does not provide for the subrange types so useful in this regard)
typedef unsigned char BYTES;
BYTES Mem[MemSize];

Since the memory is used to store not only "data" but also "instructions", another important internal
register in a processor, the so-called program counter or instruction pointer (denoted by PC or
IP), is used to keep track of the address in memory of the next instruction to be fed to the
processor’s instruction register (IR).

Perhaps it will be helpful to think of the processor itself in high-level terms:

TYPE
PROCESSOR = struct processor {
RECORD BYTES IR;
IR, BYTES R1, R2, R3;
R1, R2, R3 : BYTES; unsigned PC;
PC : ADDRESS; };
END;
VAR processor cpu;
CPU : PROCESSOR;

The operation of the machine is repeatedly to fetch a byte at a time from memory (along the data
bus), place it in the IR, and then execute the operation which this byte represents. Multi-byte
instructions may require the fetching of further bytes before the instruction itself can be decoded
fully by the CPU, of course. After the instruction denoted by the contents of IR has been executed,
the value of PC will have been changed to point to the next instruction to be fetched. This
fetch-execute cycle may be described by the following algorithm:
BEGIN
CPU.PC := initialValue; (* address of first code instruction *)
LOOP
CPU.IR := Mem[CPU.PC]; (* fetch *)
Increment(CPU.PC); (* bump PC in anticipation *)
Execute(CPU.IR); (* affecting other registers, memory, PC *)
(* handle machine interrupts if necessary *)
END
END.

Normally the value of PC alters by small steps (since instructions are usually stored in memory in
sequence); execution of branch instructions may, however, have a rather more dramatic effect. So
might the occurrence of hardware interrupts, although we shall not discuss interrupt handling
further.

A program for such a machine consists, in the last resort, of a long string of byte values. Were these
to be written on paper (as binary, decimal, or hexadecimal values), they would appear pretty
meaningless to the human reader. We might, for example, find a section of program reading
25 45 21 34 34 30 45

Although it may not be obvious, this might be equivalent to a high-level statement like
Price := 2 * Price + MarkUp;

Machine-level programming is usually performed by associating mnemonics with the recognizable

operations, like HLT for "halt" or ADD for "add to register". The above code is far more
comprehensible when written (with commentary) as
LDA 45 ; load accumulator with value stored in memory location 45
SHL ; shift accumulator one bit left (multiply by 2)
ADI 34 ; add 34 to the accumulator
STA 45 ; store the value in the accumulator at memory location 45

Programs written in an assembly language - which have first to be assembled before they can be
executed - usually make use of other named entities, for example
MarkUp EQU 34 ; CONST MarkUp = 34;
LDA Price ; CPU.A := Price;
SHL ; CPU.A := 2 * CPU.A;
ADI MarkUp ; CPU.A := CPU.A + 34;
STA Price ; Price := CPU.A;

When we use code fragments such as these for illustration we shall make frequent use of
commentary showing an equivalent fragment written in a high-level language. Commentary follows
the semicolon on each line, a common convention in assembler languages.

4.2 Addressing modes

As the examples given earlier suggest, programs prepared at or near the machine level frequently
consist of a sequence of simple instructions, each involving a machine-level operation and one or
more parameters.

An example of a simple operation expressed in a high-level language might be

AmountDue := Price + Tax;

Some machines and assembler languages provide for such operations in terms of so-called
three-address code, in which an operation - denoted by a mnemonic usually called the opcode - is
followed by two operands and a destination. In general this takes the form
operation destination, operand 1, operand 2

for example
ADD AmountDue, Price, Tax

We may also express this in a general sense as a function call

destination := operation(operand 1, operand 2 )

which helps to stress the important idea that the operands really denote "values", while the
destination denotes a processor register, or an address in memory where the result is to be stored.

In many cases this generality is restricted (that is, the machine suffers from non-orthogonality in
design). Typically the value of one operand is required to be the value originally stored at the
destination. This corresponds to high-level statements like
Price := Price * InflationFactor;

and is mirrored at the low-level by so-called two-address code of the general form
operation destination, operand
for example
MUL Price, InflationFactor

In passing, we should point out an obvious connection between some of the assignment operations
in C++ and two-address code. In C++ the above assignment would probably have been written
Price *= InflationFactor;

which, while less transparent to a Modula-2 programmer, is surely a hint to a C++ compiler to
generate code of this form. (Perhaps this example may help you understand why C++ is regarded by
some as the world’s finest assembly language!)

In many real machines even general two-address code is not found at the machine level. One of
destination and operand might be restricted to denoting a machine register (the other one might
denote a machine register, or a constant, or a machine address). This is often called one and a half
address code, and is exemplified by
MOV R1, Value ; CPU.R1 := Value
ADD Answer, R1 ; Answer := Answer + CPU.R1
MOV Result, R2 ; Result := CPU.R2

Finally, in so-called accumulator machines we may be restricted to one-address code, where the
destination is always a machine register (except for those operations that copy (store) the contents
of a machine register into memory). In some assembler languages such instructions may still appear
to be of the two-address form, as above. Alternatively they might be written in terms of opcodes
that have the register implicit in the mnemonic, for example
LDA Value ; CPU.A := Value
ADA Answer ; CPU.A := CPU.A + Answer
STB Result ; Result := CPU.B

Although many of these examples might give the impression that the corresponding machine level
operations require multiple bytes for their representation, this is not necessarily true. For example,
operations that only involve machine registers, exemplified by
MOV R1, R2 ; CPU.R1 := CPU.R2
LDA B ; CPU.A := CPU.B
TAX ; CPU.X := CPU.A

might require only a single byte - as would be most obvious in an assembler language that used the
third representation. The assembly of such programs is be eased considerably by a simple and
self-consistent notation for the source code, a subject that we shall consider further in a later
chapter.

In those instructions that do involve the manipulation of values other than those in the machine
registers alone, multi-byte instructions are usually required. The first byte typically specifies the
operation itself (and possibly the register or registers that are involved), while the remaining bytes
specify the other values (or the memory addresses of the other values) involved. In such
instructions there are several ways in which the ancillary bytes might be used. This variety gives
rise to what are known as different addressing modes for the processor, and whose purpose it is to
provide an effective address to be used in an instruction. Exactly which modes are available varies
tremendously from processor to processor, and we can mention only a few representative examples
here. The various possibilities may be distinguished in some assembler languages by the use of
different mnemonics for what at first sight appear to be closely related operations. In other
assembler languages the distinction may be drawn by different syntactic forms used to specify the
registers, addresses or values. One may even find different assembler languages for a common
processor.

In inherent addressing the operand is implicit in the opcode itself, and often the instruction is
contained in a single byte. For example, to clear a machine register named A we might have
CLA or CLR A ; CPU.A := 0

Again we stress that, though the second form seems to have two components, it does not always
imply the use of two bytes of code at the machine level.

In immediate addressing the ancillary bytes for an instruction typically give the actual value that
is to be combined with a value in a register. Examples might be
ADI 34 or ADD A, #34 ; CPU.A := CPU.A + 34

In these two addressing modes the use of the word "address" is almost misleading, as the value of
the ancillary bytes may often have nothing to do with a memory address at all. In the modes now to
be discussed the connection with memory addresses is far more obvious.

In direct or absolute addressing the ancillary bytes typically specify the memory address of the
value that is to be retrieved or combined with the value in a register, or specify where a register
value is to be stored. Examples are
LDA 34 or MOV A, 34 ; CPU.A := Mem[34]
STA 45 MOV 45, A ; Mem[45] := CPU.A
ADD 38 ADD A, 38 ; CPU.A := CPU.A + Mem[38]

Beginners frequently confuse immediate and direct addressing, a situation not improved by the fact
that there is no consistency in notation between different assembler languages, and there may even
be a variety of ways of expressing a particular addressing mode. For example, for the Intel 80x86
processors as used in the IBM-PC and compatibles, low-level code is written in a two-address form
similar to that shown above - but the immediate mode is denoted without needing a special symbol
like #, while the direct mode may have the address in brackets:
ADD AX, 34 ; CPU.AX := CPU.AX + 34 Immediate
MOV AX, [34] ; CPU.AX := Mem[34] Direct

In register-indexed addressing one of the operands in an instruction specifies both an address and
also an index register, whose value at the time of execution may be thought of as specifying the
subscript to an array stored from that address
LDX 34 or MOV A, 34[X] ; CPU.A := Mem[34 + CPU.X]
STX 45 MOV 45[X], A ; Mem[45+CPU.X] := CPU.A
ADX 38 ADD A, 38[X] ; CPU.A := CPU.A + Mem[38+CPU.X]

In register-indirect addressing one of the operands in an instruction specifies a register whose

value at the time of execution gives the effective address where the value of the operand is to be
found. This relates to the concept of pointers as used in Modula-2, Pascal and C++.
MOV R1, @R2 ; CPU.R1 := Mem[CPU.R2]
MOV AX, [BX] ; CPU.AX := Mem[CPU.BX]

Not all the registers in a machine can necessarily be used in these ways. Indeed, some machines
have rather awkward restrictions in this regard.

Some processors allow for very powerful variations on indexed and indirect addressing modes. For
example, in memory-indexed addressing, a single operand may specify two memory addresses -
the first of which gives the address of the first element of an array, and the second of which gives
the address of a variable whose value will be used as a subscript to the array.
MOV R1, 400[100] ; CPU.R1 := Mem[400 + Mem[100]]

Similarly, in memory-indirect addressing one of the operands in an instruction specifies a

memory address at which will be found a value that forms the effective address where another
operand is to be found.
MOV R1, @100 ; CPU.R1 := Mem[Mem[100]]

This mode is not as commonly found as the others; where it does occur it directly corresponds to
the use of pointer variables in languages that support them. Code like
TYPE
ARROW = POINTER TO CARDINAL; typedef int *ARROW;
VAR
Arrow : ARROW; ARROW Arrow;
Target : CARDINAL; int Target;
BEGIN
Target := Arrow^; Target = *Arrow;

might translate to equivalent code in assembler like

MOV AX, @Arrow
MOV Target, AX

or even
MOV Target, @Arrow

where, once again, we can see an immediate correspondence between the syntax in C++ and the
corresponding assembler.

Finally, in relative addressing an operand specifies an amount by which the current program count
register PC must be incremented or decremented to find the actual address of interest. This is
chiefly found in "branching" instructions, rather than in those that move data between various
registers and/or locations in memory.

Further reading

Most books on assembler level programming have far deeper discussions of the subject of
addressing modes than we have presented. Two very readable accounts are to be found in the books
by Wakerly (1981) and MacCabe (1993). A deeper discussion of machine architectures is to be
found in the book by Hennessy and Patterson (1990).

4.3 Case study 1 - A single-accumulator machine

Although sophisticated processors may have several registers, their basic principles - especially as
they apply to emulation - may be illustrated by the following model of a single-accumulator
processor and computer, very similar to one suggested by Wakerly (1981). Here we shall take
things to extremes and presume the existence of a system with all registers only 1 byte (8 bits)
wide.
4.3.1 Machine architecture

Diagrammatically we might represent this machine as in Figure 4.2.

The symbols in this diagram refer to the following components of the machine

ALU is the arithmetic logic unit, where arithmetic and logical operations are actually
performed.

A is the 8-bit accumulator, a register for doing arithmetic or logical operations.

SP is an 8-bit stack pointer, a register that points to an area in memory that may be
utilized as a stack.

X is an 8-bit index register, which is used in indexing areas of memory which

conceptually form data arrays.

Z, P, C are single bit condition flags or status registers, which are set "true" when an
operation causes a register to change to a zero value, or to a positive value, or to
propagate a carry, respectively.

IR is the 8-bit instruction register, in which is held the byte value of the instruction
currently being executed.

PC is the 8-bit program counter, which contains the address in memory of the
instruction that is next to be executed.

EAR is the effective address register, which contains the address of the byte of data
which is being manipulated by the current instruction.

The programmer’s model of this sort of machine is somewhat simpler - it consists of a number of
"variables" (in the C++ or Modula-2 sense), each of which is one byte in capacity. Some of these
correspond to processor registers, while the others form the random access read/write (RAM)
memory, of which we have assumed there to be 256 bytes, addressed by the values 0 through 255.
In this memory, as usual, will be stored both the data and the instructions for the program under
execution. The processor, its registers, and the associated RAM memory can be thought of as
though they were described by declarations like
TYPE
BYTES = CARDINAL [0 .. 255]; typedef unsigned char bytes;
PROCESSOR = RECORD struct processor {
A, SP, X, IR, PC : BYTES; bytes a, sp, x, ir, pc;
Z, P, C : BOOLEAN; bool z, p, c;
END; };
 TYPE STATUS = (running, finished, typedef enum { running, finished,
nodata, baddata, nodata, baddata, badop
badop); } status;
VAR
CPU : PROCESSOR; processor cpu;
Mem : ARRAY BYTES OF BYTES; bytes mem[256];
PS : STATUS; status ps;

where the concept of the processor status PS has been introduced in terms of an enumeration that
defines the states in which an emulator might find itself.

4.3.2 Instruction set

Some machine operations are described by a single byte. Others require two bytes, and have the
format
Byte 1 Opcode
Byte 2 Address field

The set of machine code functions available is quite small. Those marked * affect the P and Z flags,
and those marked + affect the C flag. An informal description of their semantics follows:

Mnemonic Hex Decimal Function

opcode

NOP 00h 0 No operation (this might be used to set a break point in an emulator)
CLA 01h 1 Clear accumulator A
CLC + 02h 2 Clear carry bit C
CLX 03h 3 Clear index register X
CMC + 04h 4 Complement carry bit C
INC * 05h 5 Increment accumulator A by 1
DEC * 06h 6 Decrement accumulator A by 1
INX * 07h 7 Increment index register X by 1
DEX * 08h 8 Decrement index register X by 1
TAX 09h 9 Transfer accumulator A to index register X
INI * 0Ah 10 Load accumulator A with integer read from input in decimal
INH * 0Bh 11 Load accumulator A with integer read from input in hexadecimal
INB * 0Ch 12 Load accumulator A with integer read from input in binary
INA * 0Dh 13 Load accumulator A with ASCII value read from input (a single character)
OTI 0Eh 14 Write value of accumulator A to output as a signed decimal number
OTC 0Fh 15 Write value of accumulator A to output as an unsigned decimal number
OTH 10h 16 Write value of accumulator A to output as an unsigned hexadecimal number
OTB 11h 17 Write value of accumulator A to output as an unsigned binary number
OTA 12h 18 Write value of accumulator A to output as a single character
PSH 13h 19 Decrement SP and push value of accumulator A onto stack
POP * 14h 20 Pop stack into accumulator A and increment SP
SHL + * 15h 21 Shift accumulator A one bit left
SHR + * 16h 22 Shift accumulator A one bit right
RET 17h 23 Return from subroutine (return address popped from stack)
HLT 18h 24 Halt program execution

The above are all single-byte instructions. The following are all double-byte instructions.

LDA B * 19h 25 Load accumulator A directly with contents of location whose address is
given as B
LDX B * 1Ah 26 Load accumulator A with contents of location whose address is given as B,
indexed by the value of X (that is, an address computed as the value of B + X)
LDI B * 1Bh 27 Load accumulator A with the immediate value B
LSP B 1Ch 28 Load stack pointer SP with contents of location whose address is given as B
LSI B 1Dh 29 Load stack pointer SP immediately with the value B
STA B 1Eh 30 Store accumulator A on the location whose address is given as B
 STX B 1Fh 31 Store accumulator A on the location whose address is given as B, indexed
by the value of X
ADD B + * 20h 32 Add to accumulator A the contents of the location whose address is given as B
ADX B + * 21h 33 Add to accumulator A the contents of the location whose address is given as
B,indexed by the value of X
ADI B + * 22h 34 Add the immediate value B to accumulator A
ADC B + * 23h 35 Add to accumulator A the value of the carry bit C plus the contents of the
location whose address is given as B
ACX B + * 24h 36 Add to accumulator A the value of the carry bit C plus the contents of the
location whose address is given as B, indexed by the value of X
ACI B + * 25h 37 Add the immediate value B plus the value of the carry bit C to accumulator A
SUB B + * 26h 38 Subtract from accumulator A the contents of the location whose address is
given as B
SBX B + * 27h 39 Subtract from accumulator A the contents of the location whose address is
given as B, indexed by the value of X
SBI B + * 28h 40 Subtract the immediate value B from accumulator A
SBC B + * 29h 41 Subtract from accumulator A the value of the carry bit C plus the contents
of the location whose address is given as B
SCX B + * 2Ah 42 Subtract from accumulator A the value of the carry bit C plus the contents
of the location whose address is given as B, indexed by the value of X
SCI B + * 2Bh 43 Subtract the immediate value B plus the value of the carry bit C from
accumulator A
CMP B + * 2Ch 44 Compare accumulator A with the contents of the location whose address is
given as B
CPX B + * 2Dh 45 Compare accumulator A with the contents of the location whose address is
given as B, indexed by the value of X
CPI B + * 2Eh 46 Compare accumulator A directly with the value B

These comparisons are done by virtual subtraction of the operand from A, and setting the flags P
and Z as appropriate

ANA B + * 2Fh 47 Bitwise AND accumulator A with the contents of the location whose address
is given as B
ANX B + * 30h 48 Bitwise AND accumulator A with the contents of the location whose address
is given as B, indexed by the value of X
ANI B + * 31h 49 Bitwise AND accumulator A with the immediate value B
ORA B + * 32h 50 Bitwise OR accumulator A with the contents of the location whose address
is given as B
ORX B + * 33h 51 Bitwise OR accumulator A with the contents of the location whose address
is given as B, indexed by the value of X
ORI B + * 34h 52 Bitwise OR accumulator A with the immediate value B

BRN B 35h 53 Branch to the address given as B

BZE B 36h 54 Branch to the address given as B if the Z condition flag is set
BNZ B 37h 55 Branch to the address given as B if the Z condition flag is unset
BPZ B 38h 56 Branch to the address given as B if the P condition flag is set
BNG B 39h 57 Branch to the address given as B if the P condition flag is unset
BCC B 3Ah 58 Branch to the address given as B if the C condition flag is unset
BCS B 3Bh 59 Branch to the address given as B if the C condition flag is set

JSR B 3Ch 60 Call subroutine whose address is B, pushing return address onto the stack

Most of the operations listed above are typical of those found in real machines. Notable exceptions
are provided by the I/O (input/output) operations. Most real machines have extremely primitive
facilities for doing anything like this directly, but for the purposes of this discussion we shall cheat
somewhat and assume that our machine has several very powerful single-byte opcodes for handling
I/O. (Actually this is not cheating too much, for some macro-assemblers allow instructions like this
which are converted into procedure calls into part of an underlying operating system, stored perhaps
in a ROM BIOS).
A careful examination of the machine and its instruction set will show some features that are
typical of real machines. Although there are three data registers, A, X and SP, two of them (X and
SP) can only be used in very specialized ways. For example, it is possible to transfer a value from A
to X, but not vice versa, and while it is possible to load a value into SP it is not possible to examine
the value of SP at a later stage. The logical operations affect the carry bit (they all unset it), but,
surprisingly, the INC and DEC operations do not.

It is this model upon which we shall build an emulator in section 4.3.4. In a sense the formal
semantics of these opcodes are then embodied directly in the operational semantics of the machine
(or pseudo-machine) responsible for executing them.

Exercises

4.1 Which addressing mode is used in each of the operations defined above? Which addressing
modes are not represented?

4.2 Many 8-bit microprocessors have 2-byte (16-bit) index registers, and one, two, and three-byte
instructions (and even longer). What peculiar or restrictive features does our machine possess,
compared to such processors?

4.3 As we have already commented, informal descriptions in English, as we have above, are not as
precise as semantics that are formulated mathematically. Compare the informal description of the
INC operation with the following:

INC * 05h 5 A := (A + 1) mod 256; Z := A = 0; P := A IN {0 ... 127}

Try to express the semantics of each of the other machine instructions in a similar way.

4.3.3 A specimen program

Some examples of code for this machine may help the reader’s understanding. Consider the
problem of reading a number and then counting the number of non-zero bits in its binary
representation.

Example 4.1

The listing below shows a program to solve this problem coded in an ASSEMBLER language
based on the mnemonics given previously, as it might be listed by an assembler program, showing
the hexadecimal representation of each byte and where it is located in memory.
00 BEG ; Count the bits in a number
00 0A INI ; Read(A)
01 LOOP ; REPEAT
01 16 SHR ; A := A DIV 2
02 3A 0D BCC EVEN ; IF A MOD 2 # 0 THEN
04 1E 13 STA TEMP ; TEMP := A
06 19 14 LDA BITS
08 05 INC
09 1E 14 STA BITS ; BITS := BITS + 1
0B 19 13 LDA TEMP ; A := TEMP
0D 37 01 EVEN BNZ LOOP ; UNTIL A = 0
0F 19 14 LDA BITS ;
11 0E OTI ; Write(BITS)
12 18 HLT ; terminate execution
13 TEMP DS 1 ; VAR TEMP : BYTE
14 00 BITS DC 0 ; BITS : BYTE
 15 END

Example 4.2 (absolute byte values)

In a later chapter we shall discuss how this same program can be translated into the following
corresponding absolute format (expressed this time as decimal numbers):
10 22 58 13 30 19 25 20 5 30 20 25 19 55 1 25 20 14 24 0 0

Example 4.3 (mnemonics with absolute address fields)

For the moment, we shall allow ourselves to consider the absolute form as equivalent to a form in
which the mnemonics still appear for the sake of clarity, but where the operands have all been
converted into absolute (decimal) addresses and values:
INI
SHR
BCC 13
STA 19
LDA 20
INC
STA 20
LDA 19
BNZ 1
LDA 20
OTI
HLT
0
0

Exercises

4.4 The machine does not possess an instruction for negating the value in the accumulator. What
code would one have to write to be able to achieve this?

4.5 Similarly, it does not possess instructions for multiplication and division. Is it possible to use
the existing instructions to develop code for doing these operations? If so, how efficiently can they
be done?

4.6 Try to write programs for this machine that will

(a) Find the largest of three numbers.

(b) Find the largest and the smallest of a list of numbers terminated by a zero (which is
not regarded as a member of the list).

(c) Find the average of a list of non-zero numbers, the list being terminated by a zero.

(d) Compute N! for small N. Try using an iterative as well as a recursive approach.

(e) Read a word and then write it backwards. The word is terminated with a period. Try
using an "array", or alternatively, the "stack".

(f) Determine the prime numbers between 0 and 255.

(g) Determine the longest repeated sequence in a sequence of digits terminated with
 zero. For example, for data reading 1 2 3 3 3 3 4 5 4 4 4 4 4 4 4 6 5 5 report that "4
appeared 7 times".

(h) Read an input sequence of numbers terminated with zero, and then extract the
embedded monotonically increasing sequence. For example, from 1 2 12 7 4 14 6 23
extract the sequence 1 2 12 14 23.

(i) Read a small array of integers or characters and sort them into order.

(j) Search for and report on the largest byte in the program code itself.

(k) Search for and report on the largest byte currently in memory.

(l) Read a piece of text terminated with a period, and then report on how many times
each letter appeared. To make things interesting, ignore the difference between upper
and lower case.

(m) Repeat some of the above problems using 16-bit arithmetic (storing values as pairs
of bytes, and using the "carry" operations to perform extended arithmetic).

4.7 Based on your experiences with Exercise 4.6, comment on the usefulness, redundancy and any
other features of the code set for the machine.

4.3.4 An emulator for the single-accumulator machine

Although a processor for our machine almost certainly does not exist "in silicon", its action may
easily be simulated "in software". Essentially we need only to write an emulator that models the
fetch-execute cycle of the machine, and we can do this in any suitable language for which we
already have a compiler on a real machine.

Languages like Modula-2 or C++ are highly suited to this purpose. Not only do they have
"bit-twiddling" capabilities for performing operations like "bitwise and", they have the advantage
that one can implement the various phases of translators and emulators as coherent, clearly
separated modules (in Modula-2) or classes (in C++). Extended versions of Pascal, such as Turbo
Pascal, also provide support for such modules in the form of units. C is also very suitable on the
first score, but is less well equipped to deal with clearly separated modules, as the header file
mechanism used in C is less watertight than the mechanisms in the other languages.

In modelling our hypothetical machine in Modula-2 or C++ it will thus be convenient to define an
interface in the usual way by means of a definition module, or by the public interface to a class. (In
this text we shall illustrate code in C++; equivalent code in Modula-2 and Turbo Pascal will be
found on the diskette that accompanies the book.)

The main responsibility of the interface is to declare an emulator routine for interpreting the code
stored in the memory of the machine. For expediency we choose to extend the interface to expose
the values of the operations, and the memory itself, and to provide various other useful facilities
that will help us develop an assembler or compiler for the machine in due course. (In this, and in
other interfaces, "private" members are not shown.)
// machine instructions - order is significant
enum MC_opcodes {
MC_nop, MC_cla, MC_clc, MC_clx, MC_cmc, MC_inc, MC_dec, MC_inx, MC_dex,
MC_tax, MC_ini, MC_inh, MC_inb, MC_ina, MC_oti, MC_otc, MC_oth, MC_otb,
Another Random Scribd Document
with Unrelated Content
advantage of our absence to resume the offensive. I asked him to
reduce this to writing, which he did, and I here introduce it as part
of my report:

HEADQUARTERS OF THE OHIO KNOXVILLE, December

7, 1863

Major-General W. T. SHERMAN, commanding, etc.

GENERAL: I desire to express to you and your

command my most hearty thanks and gratitude for
your promptness in coming to our relief during the
siege of Knoxville, and I am satisfied your approach
served to raise the siege. The emergency having
passed, I do not deem, for the present, any other
portion of your command but the corps of General
Granger necessary for operations in this section; and,
inasmuch as General Grant has weakened the forces
immediately with him in order to relieve us (thereby
rendering the position of General Thomas less secure),
I deem it advisable that all the troops now here, save
those commanded by General Granger, should return
at once to within supporting distance of the forces in
front of Bragg's army. In behalf of my command, I
desire again to thank you and your command for the
kindness you have done us.

I am, general, very respectfully, your obedient servant,

A. E. BURNSIDE, Major-General commanding.

Accordingly, having seen General Burnside's forces
move out of Knoxville in pursuit of Longstreet, and
General Granger's move in, I put in motion my own
command to return. General Howard was ordered to
move, via Davis's Ford and Sweetwater, to Athena,
with a guard forward at Charleston, to hold and repair
the bridge which the enemy had retaken after our
passage up. General Jeff. C. Davis moved to Columbus,
on the Hiawaesee, via Madisonville, and the two
divisions of the Fifteenth Corps moved to Tellico Plains,
to cover movement of cavalry across the mountains
into Georgia, to overtake a wagon-train which had
dodged us on our way up, and had escaped by way of
Murphy. Subsequently, on a report from General
Howard that the enemy held Charleston, I diverted
General Ewing's division to Athena, and went in person
to Tellico with General Morgan L. Smith's division. By
the 9th all our troops were in position, and we held the
rich country between the Little Tennessee and the
Hiawasaee. The cavalry, under Colonel Long, passed
the mountain at Tellico, and proceeded about
seventeen miles beyond Murphy, when Colonel Long,
deeming his farther pursuit of the wagon-train useless,
returned on the 12th to Tellico. I then ordered him and
the division of General Morgan L. Smith to move to
Charleston, to which point I had previously ordered the
corps of General Howard.

On the 14th of December all of my command in the

field lay along the Hiawassee. Having communicated to
General Grant the actual state of affairs, I received
orders to leave, on the line of the Hiawassee, all the
cavalry, and come to Chattanooga with the rest of my
command. I left the brigade of cavalry commanded by
Colonel Long, reenforced by the Fifth Ohio Cavalry
(Lieutenant-Colonel Heath)—the only cavalry properly
belonging to the Fifteenth Army Corps—at Charleston,
and with the remainder moved by easy marches, by
Cleveland and Tyner's Depot, into Chattanooga, where
I received in person from General Grant orders to
transfer back to their appropriate commands the corps
of General Howard and the division commanded by
General Jeff. C. Davis, and to conduct the Fifteenth
Army Corps to its new field of operations.

It will thus appear that we have been constantly in

motion since our departure from the Big Black, in
Mississippi, until the present moment. I have been
unable to receive from subordinate commanders the
usual full, detailed reports of events, and have
therefore been compelled to make up this report from
my own personal memory; but, as soon as possible,
subordinate reports will be received and duly
forwarded.

In reviewing the facts, I must do justice to the men of

my command for the patience, cheerfulness, and
courage which officers and men have displayed
throughout, in battle, on the march, and in camp. For
long periods, without regular rations or supplies of any
kind, they have marched through mud and over rocks,
sometimes barefooted, without a murmur. Without a
moment's rest after a march of over four hundred
miles, without sleep for three successive nights, we
crossed the Tennessee, fought our part of the battle of
Chattanooga, pursued the enemy out of Tennessee,
and then turned more than a hundred and twenty
miles north and compelled Longstreet to raise the
siege of Knoxville, which gave so much anxiety to the
whole country. It is hard to realize the importance of
these events without recalling the memory of the
general feeling which pervaded all minds at
Chattanooga prior to our arrival. I cannot speak of the
Fifteenth Army Corps without a seeming vanity; but as
I am no longer its commander, I assert that there is no
better body of soldiers in America than it. I wish all to
feel a just pride in its real honors.

To General Howard and his command, to General Jeff.

C. Davis and his, I am more than usually indebted for
the intelligence of commanders and fidelity of
commands. The brigade of Colonel Bushbeck,
belonging to the Eleventh Corps, which was the first to
come out of Chattanooga to my flank, fought at the
Tunnel Hill, in connection with General Ewing's
division, and displayed a courage almost amounting to
rashness. Following the enemy almost to the tunnel-
gorge, it lost many valuable lives, prominent among
them Lieutenant-Colonel Taft, spoken of as a most
gallant soldier.

In General Howard throughout I found a polished and

Christian gentleman, exhibiting the highest and most
chivalric traits of the soldier. General Davis handled his
division with artistic skill, more especially at the
moment we encountered the enemy's rear-guard, near
Graysville, at nightfall. I must award to this division the
credit of the best order during our movement through
East Tennessee, when long marches and the necessity
of foraging to the right and left gave some reason for
disordered ranks:

Inasmuch as exception may be taken to my

explanation of the temporary confusion, during the
battle of Chattanooga, of the two brigades of General
Matthias and Colonel Raum, I will here state that I saw
the whole; and attach no blame to any one. Accidents
will happen in battle, as elsewhere; and at the point
where they so manfully went to relieve the pressure on
other parts of our assaulting line, they exposed
themselves unconsciously to an enemy vastly superior
in force, and favored by the shape of the ground. Had
that enemy come out on equal terms, those brigades
would have shown their mettle, which has been tried
more than once before and stood the test of fire. They
reformed their ranks, and were ready to support
General Ewing's division in a very few minutes; and
the circumstance would have hardly called for notice
on my part, had not others reported what was seen
from Chattanooga, a distance of nearly five miles, from
where could only be seen the troops in the open field
in which this affair occurred.

I now subjoin the best report of casualties I am able to

compile from the records thus far received:

Killed; Wounded; and Missing............... 1949

No report from General Davis's division, but loss is

small.

Among the killed were some of our most valuable

officers: Colonels Putnam, Ninety-third Illinois;
O'Meara, Ninetieth Illinois; and Torrence, Thirtieth
Iowa; Lieutenant-Colonel-Taft, of the Eleventh Corps;
and Major Bushnell, Thirteenth Illinois.

Among the wounded are Brigadier-Generals Giles A.

Smith, Corse, and Matthias; Colonel Raum; Colonel
Waugelin, Twelfth Missouri; Lieutenant-Colonel
Partridge, Thirteenth Illinois; Major P. I. Welsh, Fifty-
sixth Illinois; and Major Nathan McAlla, Tenth Iowa.

Among the missing is Lieutenant-Colonel Archer,

Seventeenth Iowa.

My report is already so long, that I must forbear

mentioning acts of individual merit. These will be
recorded in the reports of division commanders, which
I will cheerfully indorse; but I must say that it is but
justice that colonels of regiments, who have so long
and so well commanded brigades, as in the following
cases, should be commissioned to the grade which
they have filled with so much usefulness and credit to
the public service, viz.: Colonel J. R. Cockerell,
Seventieth, Ohio; Colonel J. M. Loomis, Twenty-sixth
Illinois; Colonel C. C. Walcutt, Forty-sixth Ohio; Colonel
J. A. Williamson, Fourth Iowa; Colonel G. B. Raum,
Fifty-sixth Illinois; Colonel J. I. Alexander, Fifty-ninth
Indiana.

My personal staff, as usual, have served their country

with fidelity, and credit to themselves, throughout
these events, and have received my personal thanks.

Inclosed you will please find a map of that part of the

battle-field of Chattanooga fought over by the troops
under my command, surveyed and drawn by Captain
Jenney, engineer on my staff. I have the honor to be,
your obedient servant,

W. T. SHERMAN, Major-General commanding.

[General Order No. 68.]

WAR DEPARTMENT ADJUTANT-GENERAL'S OFFICE

WASHINGTON, February 21, 1884

Joint resolution tendering the thanks of Congress to

 Major-General W. T. Sherman and others.

Be it resolved by the Senate and House of

Representatives of the United States of America in
Congress assembled, That the thanks of Congress and
of the people of the United States are due, and that
the same are hereby tendered, to Major-General W. T.
Sherman, commander of the Department and Army of
the Tennessee, and the officers and soldiers who
served under him, for their gallant and arduous
services in marching to the relief of the Army of the
Cumberland, and for their gallantry and heroism in the
battle of Chattanooga, which contributed in a great
degree to the success of our arms in that glorious
victory.

Approved February 19, 1864. By order of the Secretary

of War:

E. D. TOWNSEND, Assistant Adjutant-General.

On the 19th of December I was at Bridgeport, and gave all the

orders necessary for the distribution of the four divisions of the
Fifteenth Corps along the railroad from Stevenson to Decatur, and
the part of the Sixteenth Corps; commanded by General Dodge,
along the railroad from Decatur to Nashville, to make the needed
repairs, and to be in readiness for the campaign of the succeeding
year; and on the 21st I went up to Nashville, to confer with General
Grant and conclude the arrangements for the winter. At that time
General Grant was under the impression that the next campaign
would be up the valley of East Tennessee, in the direction of
Virginia; and as it was likely to be the last and most important
campaign of the war, it became necessary to set free as many of the
old troops serving along the Mississippi River as possible. This was
the real object and purpose of the Meridian campaign, and of
Banks's expedition up Red River to Shreveport during that winter.
 CHAPTER XV.
MERIDIAN CAMPAIGN.

JANUARY AND FEBRUARY, 1864.

 Full Size

The winter of 1863-'64 opened very cold and severe; and it was
manifest after the battle of Chattanooga, November 25, 1863, and
the raising of the siege of Knoxville, December 5th, that military
operations in that quarter must in a measure cease, or be limited to
Burnside's force beyond Knoxville. On the 21st of December General
Grant had removed his headquarters to Nashville, Tennessee,
leaving General George H. Thomas at Chattanooga, in command of
the Department of the Cumberland, and of the army round about
that place; and I was at Bridgeport, with orders to distribute my
troops along the railroad from Stevenson to Decatur, Alabama, and
from Decatur up toward Nashville.
General G. M. Dodge, who was in command of the detachment of
the Sixteenth Corps, numbering about eight thousand men, had not
participated with us in the battle of Chattanooga, but had remained
at and near Pulaski, Tennessee, engaged in repairing that railroad,
as auxiliary to the main line which led from Nashville to Stevenson,
and Chattanooga. General John A. Logan had succeeded to the
command of the Fifteenth Corps, by regular appointment of the
President of the United States, and had relieved General Frank P.
Blair, who had been temporarily in command of that corps during the
Chattanooga and Knoxville movement.
At that time I was in command of the Department of the
Tennessee, which embraced substantially the territory on the east
bank of the Mississippi River, from Natchez up to the Ohio River, and
thence along the Tennessee River as high as Decatur and Bellefonte,
Alabama. General McPherson was at Vicksburg and General Hurlbut
at Memphis, and from them I had the regular reports of affairs in
that quarter of my command. The rebels still maintained a
considerable force of infantry and cavalry in the State of Mississippi,
threatening the river, whose navigation had become to us so delicate
and important a matter. Satisfied that I could check this by one or
two quick moves inland, and thereby set free a considerable body of
men held as local garrisons, I went up to Nashville and represented
the case to General Grant, who consented that I might go down the
Mississippi River, where the bulk of my command lay, and strike a
blow on the east of the river, while General Banks from New Orleans
should in like manner strike another to the west; thus preventing
any further molestation of the boats navigating the main river, and
thereby widening the gap in the Southern Confederacy.
 After having given all the necessary orders for the distribution,
during the winter months, of that part of my command which was in
Southern and Middle Tennessee, I went to Cincinnati and Lancaster,
Ohio, to spend Christmas with my family; and on my return I took
Minnie with me down to a convent at Reading, near Cincinnati,
where I left her, and took the cars for Cairo, Illinois, which I reached
January 3d, a very cold and bitter day. The ice was forming fast, and
there was great danger that the Mississippi River, would become
closed to navigation. Admiral Porter, who was at Cairo, gave me a
small gunboat (the Juliet), with which I went up to Paducah, to
inspect that place, garrisoned by a small force; commanded by
Colonel S. G. Hicks, Fortieth Illinois, who had been with me and was
severely wounded at Shiloh. Returning to Cairo, we started down the
Mississippi River, which was full of floating ice. With the utmost
difficulty we made our way through it, for hours floating in the midst
of immense cakes, that chafed and ground our boat so that at times
we were in danger of sinking. But about the 10th of January we
reached Memphis, where I found General Hurlbut, and explained to
him my purpose to collect from his garrisons and those of
McPherson about twenty thousand men, with which in February to
march out from Vicksburg as far as Meridian, break up the Mobile &
Ohio Railroad, and also the one leading from Vicksburg to Selma,
Alabama. I instructed him to select two good divisions, and to be
ready with them to go along. At Memphis I found Brigadier-General
W. Sooy Smith, with a force of about twenty-five hundred cavalry,
which he had by General Grant's orders brought across from Middle
Tennessee, to assist in our general purpose, as well as to punish the
rebel General Forrest, who had been most active in harassing our
garrisons in West Tennessee and Mississippi. After staying a couple
of days at Memphis, we continued on in the gunboat Silver Cloud to
Vicksburg, where I found General McPherson, and, giving him similar
orders, instructed him to send out spies to ascertain and bring back
timely information of the strength and location of the enemy. The
winter continued so severe that the river at Vicksburg was full of
floating ice, but in the Silver Cloud we breasted it manfully, and got
back to Memphis by the 20th. A chief part of the enterprise was to
destroy the rebel cavalry commanded by General Forrest, who were
a constant threat to our railway communications in Middle
Tennessee, and I committed this task to Brigadier-General W. Sooy
Smith. General Hurlbut had in his command about seven thousand
five hundred cavalry, scattered from Columbus, Kentucky, to Corinth,
Mississippi, and we proposed to make up an aggregate cavalry force
of about seven thousand "effective," out of these and the twenty-
five hundred which General Smith had brought with him from Middle
Tennessee. With this force General Smith was ordered to move from
Memphis straight for Meridian, Mississippi, and to start by February
1st. I explained to him personally the nature of Forrest as a man,
and of his peculiar force; told him that in his route he was sure to
encounter Forrest, who always attacked with a vehemence for which
he must be prepared, and that, after he had repelled the first attack,
he must in turn assume the most determined offensive, overwhelm
him and utterly destroy his whole force. I knew that Forrest could
not have more than four thousand cavalry, and my own movement
would give employment to every other man of the rebel army not
immediately present with him, so that he (General Smith) might
safely act on the hypothesis I have stated.
Having completed all these preparations in Memphis, being
satisfied that the cavalry force would be ready to start by the 1st of
February, and having seen General Hurlbut with his two divisions
embark in steamers for Vicksburg, I also reembarked for the same
destination on the 27th of January.
On the 1st of February we rendezvoused in Vicksburg, where I
found a spy who had been sent out two weeks before, had been to
Meridian, and brought back correct information of the state of facts
in the interior of Mississippi. Lieutenant-General (Bishop) Polk was in
chief command, with headquarters at Meridian, and had two
divisions of infantry, one of which (General Loring's) was posted at
Canton, Mississippi, the other (General French's) at Brandon. He had
also two divisions of cavalry—Armstrong's, composed of the three
brigades of Ross, Stark, and Wirt Adams, which were scattered from
the neighborhood of Yazoo City to Jackson and below; and Forrest's,
which was united, toward Memphis, with headquarters at Como.
General Polk seemed to have no suspicion of our intentions to
disturb his serenity.
Accordingly, on the morning of February 3d, we started in two
columns, each of two divisions, preceded by a light force of cavalry,
commanded by Colonel E. F. Winslow. General McPherson
commanded the right column, and General Hurlbut the left. The
former crossed the Big Black at the railroad-bridge, and the latter
seven miles above, at Messinger's. We were lightly equipped as to
wagons, and marched without deployment straight for Meridian,
distant one hundred and fifty miles. We struck the rebel cavalry
beyond the Big Black, and pushed them pell-mell into and beyond
Jackson during the 6th. The next day we reached Brandon, and on
the 9th Morton, where we perceived signs of an infantry
concentration, but the enemy did not give us battle, and retreated
before us. The rebel cavalry were all around us, so we kept our
columns compact and offered few or no chances for their dashes. As
far as Morton we had occupied two roads, but there we were forced
into one. Toward evening of the 12th, Hurlbut's column passed
through Decatur, with orders to go into camp four miles beyond at a
creek. McPherson's head of column was some four miles behind, and
I personally detached one of Hurlbut's regiments to guard the cross-
roads at Decatur till the head of McPherson's column should come in
sight. Intending to spend the night in Decatur, I went to a double
log-house, and arranged with the lady for some supper. We
unsaddled our horses, tied them to the fence inside the yard, and,
being tired, I lay down on a bed and fell asleep. Presently I heard
shouts and hallooing, and then heard pistol-shots close to the house.
My aide, Major Audenried, called me and said we were attacked by
rebel cavalry, who were all around us. I jumped up and inquired
where was the regiment of infantry I had myself posted at the cross-
roads. He said a few moments before it had marched past the
house, following the road by which General Hurlbut had gone, and I
told him to run, overtake it, and bring it back. Meantime, I went out
into the back-yard, saw wagons passing at a run down the road, and
horsemen dashing about in a cloud of dust, firing their pistols, their
shots reaching the house in which we were. Gathering the few
orderlies and clerks that were about, I was preparing to get into a
corn-crib at the back side of the lot, wherein to defend ourselves,
when I saw Audenried coming back with the regiment, on a run,
deploying forward as they came. This regiment soon cleared the
place and drove the rebel cavalry back toward the south, whence
they had come.
It transpired that the colonel of this infantry regiment, whose
name I do not recall, had seen some officers of McPherson's staff
(among them Inspector-General Strong) coming up the road at a
gallop, raising a cloud of duet; supposing them to be the head of
McPherson's column, and being anxious to get into camp before
dark, he had called in his pickets and started down the road, leaving
me perfectly exposed. Some straggling wagons, escorted by a New
Jersey regiment, were passing at the time, and composed the rear
of Hurlbut's train. The rebel cavalry, seeing the road clear of troops,
and these wagons passing, struck them in flank, shot down the
mules of three or four wagons, broke the column, and began a
general skirmish. The escort defended their wagons as well as they
could, and thus diverted their attention; otherwise I would surely
have been captured. In a short time the head of McPherson's
column came up, went into camp, and we spent the night in
Decatur.
The next day we pushed on, and on the 14th entered Meridian,
the enemy retreating before us toward Demopolis, Alabama. We at
once set to work to destroy an arsenal, immense storehouses, and
the railroad in every direction. We staid in Meridian five days,
expecting every hour to hear of General Sooy Smith, but could get
no tidings of him whatever. A large force of infantry was kept at
work all the time in breaking up the Mobile & Ohio Railroad south
and north; also the Jackson & Selma Railroad, east and west. I was
determined to damage these roads so that they could not be used
again for hostile purposes during the rest of the war. I never had the
remotest idea of going to Mobile, but had purposely given out that
idea to the people of the country, so as to deceive the enemy and to
divert their attention. Many persons still insist that, because we did
not go to Mobile on this occasion, I had failed; but in the following
letter to General Banks, of January 31st, written from Vicksburg
before starting for Meridian, it will be seen clearly that I indicated
my intention to keep up the delusion of an attack on Mobile by land,
whereas I promised him to be back to Vicksburg by the 1st of March,
so as to cooperate with him in his contemplated attack on
Shreveport:

HEADQUARTERS DEPARTMENT OF THE TENNESSEE

VICKSBURG, January 31, 1864

Major-General N. P. BANKS, commanding Department

of the Gulf, New Orleans.

GENERAL: I received yesterday, at the hands of

Captain Durham, aide-de-camp, your letter of the 25th
inst., and hasten to reply. Captain Durham has gone to
the mouth of White River, en route for Little Rock, and
the other officers who accompanied him have gone up
to Cairo, as I understand, to charter twenty-five
steamboats for the Red River trip. The Mississippi
River, though low for the season, is free of ice and in
good boating order; but I understand that Red River is
still low. I had a man in from Alexandria yesterday,
who reported the falls or rapids at that place
impassable save by the smallest boats. My inland
expedition is now moving, and I will be off for Jackson
and Meridian to-morrow. The only fear I have is in the
weather. All the other combinations are good. I want
to keep up the delusion of an attack on Mobile and the
Alabama River, and therefore would be obliged if you
would keep up an irritating foraging or other
expedition in that direction.

My orders from General Grant will not, as yet, justify

me in embarking for Red River, though I am very
anxious to move in that direction. The moment I
learned that you were preparing for it, I sent a
communication to Admiral Porter, and dispatched to
General Grant at Chattanooga, asking if he wanted me
and Steele to cooperate with you against Shreveport;
and I will have his answer in time, for you cannot do
any thing till Red River has twelve feet of water on the
rapids at Alexandria. That will be from March to June. I
have lived on Red River, and know somewhat of the
phases of that stream. The expedition on Shreveport
should be made rapidly, with simultaneous movements
from Little Rock on Shreveport, from Opelousas on
Alexandria, and a combined force of gunboats and
transports directly up Red River. Admiral Porter will be
able to have a splendid fleet by March 1st. I think
Steele could move with ten thousand infantry and five
thousand cavalry. I could take about ten thousand, and
you could, I suppose, have the same. Your movement
from Opelousas, simultaneous with mine up the river,
would compel Dick Taylor to leave Fort De Russy (near
Marksville), and the whole combined force could
appear at Shreveport about a day appointed
beforehand.

I doubt if the enemy will risk a siege at Shreveport,

although I am informed they are fortifying the place,
and placing many heavy guns in position. It would be
better for us that they should stand there, as we might
make large and important captures. But I do not
believe the enemy will fight a force of thirty thousand
men, acting in concert with gunboats.
 I will be most happy to take part in the proposed
expedition, and hope, before you have made your final
dispositions, that I will have the necessary permission.
Half the Army of the Tennessee is near the Tennessee
River, beyond Huntsville, Alabama, awaiting the
completion of the railroad, and, by present orders, I
will be compelled to hasten there to command it in
person, unless meantime General Grant modifies the
plan. I have now in this department only the force left
to hold the river and the posts, and I am seriously
embarrassed by the promises made the veteran
volunteers for furlough. I think, by March 1st, I can
put afloat for Shreveport ten thousand men, provided I
succeed in my present movement in cleaning out the
State of Mississippi, and in breaking up the railroads
about Meridian.

I am, with great respect, your obedient servant,

W. T. SHERMAN, Major-General, commanding.

The object of the Meridian expedition was to strike the roads

inland, so to paralyze the rebel forces that we could take from the
defense of the Mississippi River the equivalent of a corps of twenty
thousand men, to be used in the next Georgia campaign; and this
was actually done. At the same time, I wanted to destroy General
Forrest, who, with an irregular force of cavalry, was constantly
threatening Memphis and the river above, as well as our routes of
supply in Middle Tennessee. In this we failed utterly, because
General W. Sooy Smith did not fulfill his orders, which were clear and
specific, as contained in my letter of instructions to him of January
27th, at Memphis, and my personal explanations to him at the same
time. Instead of starting at the date ordered, February 1st, he did
not leave Memphis till the 11th, waiting for Warings brigade that was
ice-bound near Columbus, Kentucky; and then, when he did start, he
allowed General Forrest to head him off and to defeat him with an
inferior force, near West Point, below Okalona, on the Mobile & Ohio
Railroad.
We waited at Meridian till the 20th to hear from General Smith,
but hearing nothing whatever, and having utterly destroyed the
railroads in and around that junction, I ordered General McPherson
to move back slowly toward Canton. With Winslow's cavalry, and
Hurlbut's infantry, I turned north to Marion, and thence to a place
called "Union," whence I dispatched the cavalry farther north to
Philadelphia and Louisville, to feel as it were for General Smith, and
then turned all the infantry columns toward Canton, Mississippi. On
the 26th we all reached Canton, but we had not heard a word of
General Smith, nor was it until some time after (at Vicksburg) that I
learned the whole truth of General Smith's movement and of his
failure. Of course I did not and could not approve of his conduct,
and I know that he yet chafes under the censure. I had set so much
store on his part of the project that I was disappointed, and so
reported officially to General Grant. General Smith never regained
my confidence as a soldier, though I still regard him as a most
accomplished gentleman and a skillful engineer. Since the close of
the war he has appealed to me to relieve him of that censure, but I
could not do it, because it would falsify history.
Having assembled all my troops in and about Canton, on the 27th
of February I left them under the command of the senior major-
general, Hurlbut, with orders to remain till about the 3d of March,
and then to come into Vicksburg leisurely; and, escorted by
Winslow's cavalry, I rode into Vicksburg on the last day of February.
There I found letters from General Grant, at Nashville, and General
Banks, at New Orleans, concerning his (General Banks's) projected
movement up Red River. I was authorized by the former to
contribute aid to General Banks for a limited time; but General Grant
insisted on my returning in person to my own command about
Huntsville, Alabama, as soon as possible, to prepare for the spring
campaign.
About this time we were much embarrassed by a general order of
the War Department, promising a thirty-days furlough to all soldiers
who would "veteranize"—viz., reenlist for the rest of the war. This
was a judicious and wise measure, because it doubtless secured the
services of a very large portion of the men who had almost
completed a three-years enlistment, and were therefore veteran
soldiers in feeling and in habit. But to furlough so many of our men
at that instant of time was like disbanding an army in the very midst
of battle.
In order to come to a perfect understanding with General Banks, I
took the steamer Diana and ran down to New Orleans to see him.
Among the many letters which I found in Vicksburg on my return
from Meridian was one from Captain D. F. Boyd, of Louisiana, written
from the jail in Natchez, telling me that he was a prisoner of war in
our hands; had been captured in Louisiana by some of our scouts;
and he bespoke my friendly assistance. Boyd was Professor of
Ancient Languages at the Louisiana Seminary of Learning during my
administration, in 1859-'60; was an accomplished scholar, of
moderate views in politics, but, being a Virginian, was drawn, like all
others of his kind, into the vortex of the rebellion by the events of
1861, which broke up colleges and every thing at the South.
Natchez, at this time, was in my command, and was held by a
strong division, commanded by Brigadier-General J. W. Davidson. In
the Diana we stopped at Natchez, and I made a hasty inspection of
the place. I sent for Boyd, who was in good health, but quite dirty,
and begged me to take him out of prison, and to effect his
exchange. I receipted for him; took him along with me to New
Orleans; offered him money, which he declined; allowed him to go
free in the city; and obtained from General Banks a promise to effect
his exchange, which was afterward done. Boyd is now my legitimate
successor in Louisiana, viz., President of the Louisiana University,
which is the present title of what had been the Seminary of
Learning. After the war was over, Boyd went back to Alexandria,
reorganized the old institution, which I visited in 1866 but the
building was burnt down by an accident or by an incendiary about
1868, and the institution was then removed to Baton Rouge, where
it now is, under its new title of the University of Louisiana.
We reached New Orleans on the 2d of March. I found General
Banks, with his wife and daughter, living in a good house, and he
explained to me fully the position and strength of his troops, and his
plans of action for the approaching campaign. I dined with him, and,
rough as I was—just out of the woods—attended, that night, a very
pleasant party at the house of a lady, whose name I cannot recall,
but who is now the wife of Captain Arnold, Fifth United States
Artillery. At this party were also Mr. and Mrs. Frank Howe. I found
New Orleans much changed since I had been familiar with it in 1853
and in 1860-'61. It was full of officers and soldiers. Among the
former were General T. W. Sherman, who had lost a leg at Port
Hudson, and General Charles P: Stone, whom I knew so well in
California, and who is now in the Egyptian service as chief of staff.
The bulk of General Banks's army was about Opelousas, under
command of General Franklin, ready to move on Alexandria. General
Banks seemed to be all ready, but intended to delay his departure a
few days to assist in the inauguration of a civil government for
Louisiana, under Governor Hahn. In Lafayette Square I saw the
arrangements of scaffolding for the fireworks and benches for the
audience. General Banks urged me to remain over the 4th of March,
to participate in the ceremonies, which he explained would include
the performance of the "Anvil Chorus" by all the bands of his army,
and during the performance the church-bells were to be rung, and
cannons were to be fired by electricity. I regarded all such
ceremonies as out of place at a time when it seemed to me every
hour and every minute were due to the war. General Banks's
movement, however, contemplated my sending a force of ten
thousand men in boats up Red River from Vicksburg, and that a
junction should occur at Alexandria by March 17th. I therefore had
no time to wait for the grand pageant of the 4th of March, but took
my departure from New Orleans in the Diana the evening of March
3d.
On the next day, March 4th, I wrote to General Banks a letter,
which was extremely minute in conveying to him how far I felt
authorized to go under my orders from General Grant. At that time
General Grant commanded the Military Division of the Mississippi,
embracing my own Department of the Tennessee and that of
General Steele in Arkansas, but not that of General Banks in
Louisiana. General Banks was acting on his own powers, or under
the instructions of General Halleck in Washington, and our assistance
to him was designed as a loan of ten thousand men for a period of
thirty days. The instructions of March 6th to General A. J. Smith,
who commanded this detachment, were full and explicit on this
point. The Diana reached Vicksburg on the 6th, where I found that
the expeditionary army had come in from Canton. One division of
five thousand men was made up out of Hurlbut's command, and
placed under Brigadier-General T. Kilby Smith; and a similar division
was made out of McPherson's and Hurlbut's troops, and placed
under Brigadier-General Joseph A. Mower; the whole commanded by
Brigadier-General A. J. Smith. General Hurlbut, with the rest of his
command, returned to Memphis, and General McPherson remained
at Vicksburg. General A. J. Smith's command was in due season
embarked, and proceeded to Red River, which it ascended, convoyed
by Admiral Porter's fleet. General Mower's division was landed near
the outlet of the Atchafalaya, marched up by land and captured the
fort below Alexandria known as Fort De Russy, and the whole fleet
then proceeded up to Alexandria, reaching it on the day appointed,
viz., March 17th, where it waited for the arrival of General Banks,
who, however, did not come till some days after. These two divisions
participated in the whole of General Banks's unfortunate Red River
expedition, and were delayed so long up Red River, and
subsequently on the Mississippi, that they did not share with their
comrades the successes and glories of the Atlanta campaign, for
which I had designed them; and, indeed, they, did not join our army
till just in time to assist General George H. Thomas to defeat General
Hood before Nashville, on the 15th and 16th of December, 1864.
General Grant's letter of instructions, which was brought me by
General Butterfield, who had followed me to New Orleans, enjoined
on me, after concluding with General Banks the details for his Red
River expedition, to make all necessary arrangements for furloughing
the men entitled to that privilege, and to hurry back to the army at
Huntsville, Alabama. I accordingly gave the necessary orders to
General McPherson, at Vicksburg, and continued up the river toward
Memphis. On our way we met Captain Badeau, of General Grant's
staff, bearing the following letter, of March 4th, which I answered on
the 10th, and sent the answer by General Butterfield, who had
accompanied me up from New Orleans. Copies of both were also
sent to General McPherson, at Vicksburg:

[Private.]

NASHVILLE, TENNESSEE, March 4, 1864

DEAR SHERMAN: The bill reviving the grade of

lieutenant-general in the army has become a law, and
my name has been sent to the Senate for the place.

I now receive orders to report at Washington

immediately, in person, which indicates either a
confirmation or a likelihood of confirmation. I start in
the morning to comply with the order, but I shall say
very distinctly on my arrival there that I shall accept no
appointment which will require me to make that city
my headquarters. This, however, is not what I started
out to write about.
While I have been eminently successful in this war, in
at least gaining the confidence of the public, no one
feels more than I how much of this success is due to
the energy, skill, and the harmonious putting forth of
that energy and skill, of those whom it has been my
good fortune to have occupying subordinate positions
under me.

There are many officers to whom these remarks are

applicable to a greater or less degree, proportionate to
their ability as soldiers; but what I want is to express
my thanks to you and McPherson, as the men to
whom, above all others, I feel indebted for whatever I
have had of success. How far your advice and
suggestions have been of assistance, you know. How
far your execution of whatever has been given you to
do entitles you to the reward I am receiving, you
cannot know as well as I do. I feel all the gratitude this
letter would express, giving it the most flattering
construction.

The word you I use in the plural, intending it for

McPherson also. I should write to him, and will some
day, but, starting in the morning, I do not know that I
will find time just now. Your friend,

U. S. GRANT, Major-General.

[PRIVATE AND CONFIDENTIAL]

NEAR MEMPHIS, March 10, 1864

General GRANT.
DEAR GENERAL: I have your more than kind and
characteristic letter of the 4th, and will send a copy of
it to General McPherson at once.

You do yourself injustice and us too much honor in

assigning to us so large a share of the merits which
have led to your high advancement. I know you
approve the friendship I have ever professed to you,
and will permit me to continue as heretofore to
manifest it on all proper occasions.

You are now Washington's legitimate successor, and

occupy a position of almost dangerous elevation; but if
you can continue as heretofore to be yourself, simple,
honest, and unpretending, you will enjoy through life
the respect and love of friends, and the homage of
millions of human beings who will award to you a large
share for securing to them and their descendants a
government of law and stability.

I repeat, you do General McPherson and myself too

much honor. At Belmont you manifested your traits,
neither of us being near; at Donelson also you
illustrated your whole character. I was not near, and
General McPherson in too subordinate a capacity to
influence you.

Until you had won Donelson, I confess I was almost

cowed by the terrible array of anarchical elements that
presented themselves at every point; but that victory
admitted the ray of light which I have followed ever
since.

I believe you are as brave, patriotic, and just, as the

great prototype Washington; as unselfish, kind-
hearted, and honest, as a man should be; but the chief
characteristic in your nature is the simple faith in
success you have always manifested, which I can liken
to nothing else than the faith a Christian has in his
Saviour.

This faith gave you victory at Shiloh and Vicksburg.

Also, when you have completed your best
preparations, you go into battle without hesitation, as
at Chattanooga—no doubts, no reserve; and I tell you
that it was this that made us act with confidence. I
knew wherever I was that you thought of me, and if I
got in a tight place you would come—if alive.

My only points of doubt were as to your knowledge of

grand strategy, and of books of science and history;
but I confess your common-sense seems to have
supplied all this.

Now as to the future. Do not stay in Washington.

Halleck is better qualified than you are to stand the
buffets of intrigue and policy. Come out West; take to
yourself the whole Mississippi Valley; let us make it
dead-sure, and I tell you the Atlantic slope and Pacific
shores will follow its destiny as sure as the limbs of a
tree live or die with the main trunk! We have done
much; still much remains to be done. Time and time's
influences are all with us; we could almost afford to sit
still and let these influences work. Even in the seceded
States your word now would go further than a
President's proclamation, or an act of Congress.

For God's sake and for your country's sake, come out
of Washington! I foretold to General Halleck, before he
left Corinth, the inevitable result to him, and I now
exhort you to come out West. Here lies the seat of the
coming empire; and from the West, when our task is
 done, we will make short work of Charleston and
Richmond, and the impoverished coast of the Atlantic.
Your sincere friend,

W. T. SHERMAN

We reached Memphis on the 13th, where I remained some days,

but on the 14th of March received from General Grant a dispatch to
hurry to Nashville in person by the 17th, if possible. Disposing of all
matters then pending, I took a steamboat to Cairo, the cars thence
to Louisville and Nashville, reaching that place on the 17th of March,
1864.
I found General Grant there. He had been to Washington and
back, and was ordered to return East to command all the armies of
the United States, and personally the Army of the Potomac. I was to
succeed him in command of the Military Division of the Mississippi,
embracing the Departments of the Ohio, Cumberland, Tennessee,
and Arkansas. General Grant was of course very busy in winding up
all matters of business, in transferring his command to me, and in
preparing for what was manifest would be the great and closing
campaign of our civil war. Mrs. Grant and some of their children were
with him, and occupied a large house in Nashville, which was used
as an office, dwelling, and every thing combined.
On the 18th of March I had issued orders assuming command of
the Military Division of the Mississippi, and was seated in the office,
when the general came in and said they were about to present him
a sword, inviting me to come and see the ceremony. I went back
into what was the dining-room of the house; on the table lay a rose-
wood box, containing a sword, sash, spurs, etc., and round about
the table were grouped Mrs. Grant, Nelly, and one or two of the
boys. I was introduced to a large, corpulent gentleman, as the
mayor, and another citizen, who had come down from Galena to
make this presentation of a sword to their fellow-townsman. I think
Welcome to Our Bookstore - The Ultimate Destination for Book Lovers
Are you passionate about books and eager to explore new worlds of
knowledge? At our website, we offer a vast collection of books that
cater to every interest and age group. From classic literature to
specialized publications, self-help books, and children’s stories, we
have it all! Each book is a gateway to new adventures, helping you
expand your knowledge and nourish your soul
Experience Convenient and Enjoyable Book Shopping Our website is more
than just an online bookstore—it’s a bridge connecting readers to the
timeless values of culture and wisdom. With a sleek and user-friendly
interface and a smart search system, you can find your favorite books
quickly and easily. Enjoy special promotions, fast home delivery, and
a seamless shopping experience that saves you time and enhances your
love for reading.
Let us accompany you on the journey of exploring knowledge and
personal growth!

ebookball.com