BHARATI VIDYAPEETH’S

COLLEGE OF ENGINEERING, KOLHAPUR

DEPARTMENT OF COMPUTER SCIENCE AND ENGINEERING

Class: TECSE-I

SYSTEM
PROGRAMMIMG
 Unit 1: Language Processor

➢ Introduction: -
• The designer expresses the ideas in terms related to the application
domain of the software.
• To implement these ideas, their description has to be interpreted in
terms related to the execution domain of the computer system.
• The term semantics to represent the rules of the meaning of a domain
and the term semantic gap to represent the difference between the
semantics of the two domains.
• The semantic gap has many consequences. Some of the important
are once being large development efforts and poor quality of
software.
• These issues are taken by software engineering through the use of
methodologies and programming language (PL’s).

Semantic gap

Application domain Execution domain

Fig. Consequence of Semantic Gap

• The software engineering steps aimed at the use of PL can be

grouped into – a) Specification, design, coding steps.

b) PL implementation steps.

• Software implementation using a PL introduce a new domain, the

PL domain.
• The Semantic gap between application domain and the execution
domain is bridged by the software engineering steps.
• The first step bridges the gap between the application and PL
domain. While second step bridges the gap between PL and
execution domain.
• The gap between application and PL domain refers as specification
and design gap or specification gap.
• The gap between PL domain and execution domain refers as
execution gap.
• Specification gap is bridged by software development team.
• Execution gap is bridged by designer of the programming language
processor.

Fig. Specification and Execution Gap

• Specification gap: Specification gap is the gap between two

specifications of the same task.
• Execution gap: The execution gap is the gap between the semantics
of program (that performs the same tasks) but written in different
programming language.
➢ Language Processor: -
• A language processor is a software which bridges the specification
or execution gap.
• A Language Processor is a special type of computer software that
has the capacity of translating the source code or program codes into
machine codes.
• Language processing to describe the activity performed by a
language processor and assume a diagnostic capability as an implicit
part of any form of language processing.

Fig. Language Processor

• The program form input to a language processor as the source

program and its output as the target program.
• The languages in which these programs are written are called source
language and target language.

• Language translator: A Language translator bridges an execution

gap to the machine language or assembly language of computer
system. It converts a computer program from one language to
another.

Fig. Language Translator

 • Assembler: An assembler is a language translator whose source
language is assembly language. It translates from low level language
into binary machine code.

• Complier: A compiler is an any language translator which is not an

assembler. It translates source code from high level language to
binary machine code.

• De translator: A de translator bridges the same execution gap as the

language translator, but in reverse direction.

• Processor: A processor is a language processor which bridges an

execution gap. It is not a language translator.

• Migrator: A language migrator bridges the specification gap

between two PL’s. A Language Migrator makes a program
compatible to all the Operating Systems.

➢ Interpreters: -
• An interpreter is a language processor which bridges an execution
gap without generating a machine language program.
• Interpreter is a language translator.
• The absence of target program implies the absence of an output
interface of the interpreter.
• Thus the language processing activites.
• Where interpreter domain encompasses the PL domain as well as
execution domain.
 • Thus the specification language of PL domain is identical with
specification language of interpreter domain, interpreter also
incorporate with execution domain.

Fig. Schematic representation of interpreter

➢ Problem oriented language: -

• It is a program whose statements resemble terminology of the user.
• To develop a PL such that the PL domain is very close or identical
to the application domain.
• PL features now directly model aspects of the application domain,
which leads to very small specification applications. Hence they are
called as problem oriented language.
• They have large execution gaps, however this is acceptable because
the gap is bridged by the translator or interpreter does not concern
with the software designer.
 Fig. Problem Oriented Language

➢ Procedure oriented language: -

• A procedure oriented language provide general purpose facilities
required in most application domains.
• A program in procedural language is a list of instruction where each
statement tell the computer to do something.
• Such a language is independent of specific application domain and
results large specification gap which has to be bridged by an
application designer.

➢ Language Processing Activities: -

• The fundamental language processing activites can be divided into
those who bridges the specification gap and those who bridges the
execution gap.
a) Program generation activities.
b) Program execution activities.
• Program generation activity aims to generation of a program. It
bridges specification gap.
 • The source language is specification language of an application
domain and the target language is typically procedure oriented PL.
• A program execution activity aims to execute a program written in
PL. It bridges execution gap.
• It’s source language could be procedure oriented or problem
oriented.

➢ Program Generation: -
• The program generator is a software system which accepts the
specification of a program to be generated,and generates a program
in target PL.
• The program generator introduces a new domain between
application domain and PL domain i.e program generator domain.
• The application gap is the now between the application domain and
program generator domain.This gap is smaller than the gap between
the application domain and target PL domain.
• Reduction in specification gap increases the reliability of the
generated program.Since the program generator domain is close to
the application domain.
• The execution gap between the target PL domain and the execution
domain is bridged by the compiler or interpreter for the PL.

Fig. Program Generation

➢ Program Execution: -
• Two popular models for program execution are translation and
interpretation.
• Program translation: The program translation model bridges the
execution gap by translating a program written inPL,called a source
program into an equivalent program in machine or assembly
language of the computer system,called as target program.

Fig. Program Translation

• Characteristics of the program translation model:

a) A program must be translated before it can be executed.
b) The translated program may be saved in the file.The saved
program may execute repeatedly.
c) A program must be retranslated following modifications.

d) Program interpretations:

The interpreter needs a source program and stores it in

it’s memory. During interpretations it takes a source statements
determines its meaning and performs actions which implement it.
This include computational &input-output actions. The CPU uses a
program counter (PC) to note the address of next instruction to be
executed.
➢ Program Interpretation: -

Fig. Program Interpretation

e) The instruction is subjected to the instruction execution cycle

consisting of following steps:
i. Fetch the instruction.
ii. Decode the instruction to determine the operation to
be performed and also its operands.
iii. Execute the instruction.
• At the end of the cycle, the instruction address in PC is updated and
the cycle is repeated for the next instruction.
• Program interpretation can proceed in an analogous manner. Thus,
the Pc indicate which statement of the software of the source
program is to be interpreted next.
• The statement would be subjected to the interpretation cycle, which
consist following steps:
▪ The Fetch the statement.
▪ Analyze the statement &determine its meaning viz. the
computations to be performed and its operands.
▪ Execute the meaning of statement.
 • Characteristics of interpretation:
a)The source program retained in the source form itself, i.e. no
target program form exist.
b)A statement is analyzed during its interpretation.

➢ Fundamentals of language processing: -

• Language processing=
analysis of source program + Synthesis of target program
• The collection of language processor components engaged in
analyzing the source program as the analysis phase of language
processor.
• Components engaged in synthesizing a target program constitute the
synthesis phase.
• The specifications of the source language forms basis of the source
program analysis .
• The specification consist of the three components:
▪ Lexical Rules: Which governs the valid lexical units in the
source program.
▪ Syntax Rules: Which governs the formation of valid
statements in the source language.
▪ Semantic Rules: Which associates meaning with valid
statement of the language.

➢ Analysis Phase: -
• The analysis phase uses the each component of the source language
specification to determine relevant information concerning a
statement in source program.
• Thus, analysis of source statement consist of lexical,syntax,and
semantics analysis.
➢ Synthesis Phase: -
• The Synthesis phase is concerned with the construction of target
language statement which have same meaning as a source
statement.
• Typically, this consist of two main activities:
▪ Creation of data structure in target program.
▪ Generation of code.

➢ Phases and passes of language processor: -

• A language processor consist of two distinct phases-the analysis and
synthesis phase.
• Language processing can be performed on statement by statement
basis-that is analysis of source statement can be immediately
followed by synthesis of equivalent target statements.

Fig. Phases of Language Processor

• Forward Reference: A forword reference of a program entity is a

reference to the entity which precedes its definition in the program.
• While processing a statement containing a forward reference,a
language processor does not posses all relevant information
concerning the referenced entity.
• This creates the difficulties in synthesizing the equivalent target
statements.
• This problem can be solved by positioning the generation of target
code until more information concerning the entity becomes
available.
 • Postponing the generation of target code may also reduce memory
requirements of the language processor and simply its
organizations.
• Example:
Percent_profit:=(profit*100)/cost_price;
------
Long profit;
The statement long profit declares profit to have a double
precision value.The reference to profit in the assignment statement
constitute a forward reference be causes the declaration of the profit
occurs later in the program.Since type of processor is not known
while processing the assignment statement ,correcet code cannnot
be generated for it in a statement by statement manner.

➢ Language processor passes: -

• A language processor pass is the processing of every statement in
the source program,or its equivalent representation to perform a
language processing function.
• Here pass is an abstract noun describing the processing performed
by the language processor.For simplicity the part of the language
processor which performs one pass over the source program is called
as a pass.
• Pass 1:Perform analysis of the source program and note relevant
information.
• Pass 2: Perform aynthesis of target program.
• Information concerning in the pass 1,this information is used during
pass 2 to perform code generation.

➢ Intermediate representation of programs: -

• In pass 1,it analyses the source program to note thetype information.
• In pass 2,it once again analyses the source program to generate target
code using the type information noted in pass 1.
• This can be avoid using an intermediate representation of the source
program.
• Defination:An intermediate representation(IR) is a representation of
a source program which reflects the effect of some,but not
all,analysis and synthesis tasks performed during language
processing.
• The first pass performs analysis of the source program and reflects
results in intermediate representation.
• The second pass reads and analyses the IR,instead of the source
program ,to perform synthesis of the target program.This avoids
repeated processing of the source program.
• The first pass is concerned exclusively with source language
issues.Hence it is called the front end of the language processor.
• The second pass is concerned with the program synthesis for a
specific target language.Hence it is called as the back end of
language processor.
• Note that the front and back ends of the language processor need not
coexist in memory.
• This reduces the memory requirements of a language processor.

Fig. Schematic 2-Pass Language Processing

• Desirable properties of IR:

▪ Ease of use:IR should be easy to construct and
analyses.
▪ Processing efficiency:efficient algorithms must
exist for constructing and analyzing the IR.
▪ Memory efficiency:IR must be compact.
➢ A toy compiler: -
• The front end:

The front end performs lexical,syntax and semantic

analysis of the source program .Each kind of analysis involves the
following functions:

a) Determine validity of a source statement from the viewpoint of the

analysis.
b) Determine the content of a source statement.
c) Construct a suitable representation of the source statement for use
by subsequent analysis functions,or by synthesis phase of the
language processor.

Fig. A Toy Compiler

The word ‘content’ has different connotations in

lexical,syntax,and semantic analysis.In lexical analysis ,the content
is the lexical class to which each lexical unit belongs.
While in syntax analysis it is the syntactic structure of a source
statement.In semantic analysis the content is the meaning of the
statement for a declaration statement,it is the set of attributes of a
declared variables ,while for an imperative statements,it is the
sequence of action implied by the statement.Each analysis
represents the ‘content’ of a source statement in the form of tables
of information and description of the source statement.
Subsequent analysis uses this information for its own
purposes and either adds information to these tables and
descriptions,or constructs its own tables and descriptors .
The tables and descriptions at the end of the semantic analysis form
the IR of the front end.The IR produced by the front end consists of
two components:

a) Tables of information: Tables contain the

information obtained during different analyses of
SP.The most important table is the symbol table
which contains information concerning all identifires
used in SP.The symbol table is built during lexical
analysis.Sementic analysis adds information
concering symbol attributes while processing
declarations statements.It may also add new names
designing temporary results.
b) An intermediate code(IC): It is a description of the
source program.The IC is a sequence of IC units,each
IC unit representing the meaning of the one action in
SP.IC units may contain references to the information
in various tables.

Example:

IR produced by the analysis phase for the program

i=integer;
a,b=real;
a:=b+i;
 Symbol table:

Sr no. symbol type length Address

1. i int

2. a real

3. b real

4. i* real

5. temp real

Intermediate code:
1) Convert (Id,#1)to
real,giving(Id,#4).
2) Add (Id,#4),to(Id,#3),giving
(Id,#5).
3) Store (Id,#5)in (Id,#2).

• Lexical analysis(Scanning):
▪ Lexical analysis identifies the lexical units in the source
statement.
▪ It then classifies the units into different lexical classes.eg-
id’s,constants,reserved id’s etc.
▪ And enters them into different tables.
▪ Lexical analysis builds a descriptor,called a token,for each
lexical unit.A token contains a lexical unit belongs,number in
class is the entry number of the lexical unit in the relevant
table.

• Syntax analysis:
▪ Syntax analysis processes the string of tokens built by lexical
analysis to dertermine the statement class.
▪ Eg-assignment statement,if statement etc.
 ▪ It then builds an IC which represents the structure of the
statement.The IC is passed to semantic analysis to determine
the meaning of the statement.

• Semantic analysis:
▪ Semantic analysis of declaration statements differs from the
semantic analysis of imperative statements.
▪ The latter identifies the sequences of actions necessary to
implement the meaning of the source statement .
▪ In both cases the structure of a source statement guides the
application of the semantic rules.
▪ When semantic analysis determines the meaning of a subtree
in the IC,it adds information to table or adds an action to the
sequence of actions. It then modifies the IC to enable further
semantic analysis .
▪ The analysis ends when the tree has been completely
processed.
▪ The updated tables and the sequence of actions constitute the
IR produced by the analysis phase.

Fig. Front End of a Toy Compiler

• The back end:
The back end performs memory allocation and code
generation.

o Memory allocation:
▪ Memory allocation is a simple task given the presence of the
symbol table.
▪ The memory requirement of an identifier is computed from its
type,length and dimensionality,and memory is allocated ton
it.
▪ The address of the memory area is entered in the symbol table.

o Code Generation:
▪ Code generation uses the knowledge of the target
architecture,viz knowledge of instructions and addressing
modes in the target computer,to select the appropriate
instructions.
▪ The important issues in code generation are:
1) Determine the places where the intermediate results should
be kept,i.e whether they should be kept in memory
locations or held in machine registers. This is preparatory
step for code generation.
2) Determine which instructions should be used for type
conversion operators.
3) Determine which addressing modes should be used for
accessing variable.
 Fig. Back End of a Toy Compiler

➢ Fundamentals of language specification: -

Specification of the source language forms the basis of source program
analysis.

Fig. Language Specification Fundamentals

• Programming language grammars:
o The lexical and syntactic features of a programming language are
specified by its grammar.
o The language L can be considered to be a collection of valid
sentences.
o Each sentence can be looked upon as a sequence of words, and each
word as a sequence of letters or graphic symbol acceptable in L.
o A language specify in this manner is known as formal language.A
formal language grammar is a set of rules which precisely specify
the sentences of L.
o It is clear that natural languages are not formal languages because of
their rich vocabulary. However PLS are formal language.
o It includes terminal symbols, alphabets, strings, nonterminal
symbols, productions, derivation, reduction and parse trees.

• Classification of grammars:
o Grammars are defined on the basis of the nature of productions used
in them(Chomsky,1963).
o Each grammar class has its own characteristics and limitations.
o Type 0 grammars: These grammars are also known as phrase
structure grammars, contain productions of the form,
a:: =b
where both a and b can be strings of T’s and NT’s. Such productions
permits arbitrary substitution of strings during derivation or
reduction ,hence they are not relevant to specification of
programming language.

o Type 1 grammar: These grammars are known as context sensitive

grammar because their productions specify that derivation or
reduction of strings can be take place only in specific contexts. A
type -1 grammar production has the form,

a A b::=a pi b

Thus, a string pi in a sentential form can be replaced by ‘A’ only

when it is enclosed by the strings a and b. These grammars are also
 not particularly relevant for PL specification recognition of PL
constructs is not context sensitive in nature.

o Type 2 grammar: These grammars impose no context requirements

on derivation or reductions. A typical Type 2 productions is in the
form of,
A::=pi
Which can be applied independent of its context. These grammars
therefore known as context free grammar (CFG). CFG’s are ideally
suited for programming language specification.

o Type 3 grammar: Type 3 grammars are characterized by

productions of the form,
A:: tB|t or
A::=Bt|t

Note that this grammar also satisfy the type 2 grammar. Type 3
grammar are also known as linear grammars or regular grammars.

o Operator grammar: An operator grammar is a grammar none of

whose productions contain two or more consecutive NT’s in any
RHS alternative.

➢ Language Processing Development Tools: -

• The analysis phase of the language processor has a standard form
irrespective of its purpose,the source text is subjected to
lexical,syntax and semantic analysis and the result of analysis are
represented in an IR.
• LPDT focuses upon generation of the analysis phase of language
processor.
• Thus writing of language processor is a well understood and repetive
process which ideally suits the program generation approach to
software development.
• This has lead to the development of a set of language processor
development tools(LPDT’s)focusing on generation of the analysis
phase of language processor.
• The LPDT requires the following two inputs:
 o Specification of a grammar of language L.
o Specification of semantic actions to be performed in the
analysis phase.

Fig. Language Processor Development Tools (LPDT)

• It generates programs that performs lexical, syntax and semantic

analysis of the source program and construct the IR.
• These programs collectively form the analysis phase of the language
processor.
• Lexical analyzer generator LEX and the parser generator YACC.
• The input tools is a specification of the lexical and syntactic
constructs of L and the semantic actions to be performed on
recognizing the constructs.
• The specification consist of set of translation rules of the form,

<String specification> {<Semantic action> }

• Where <semantic action >consist of C code. This code is executed

when a string matching <string specification> is encountered in the
input.
• LEX and YACC generates C programs which contain the code for
scanning and parsing respectively, and the semantic actions are
contain in the specification.
• A YACC generated parser can use a LEX generated scanner as a
routine if the scanner and parser use same conventions concerning
the representation of tokens.
• Fig shows the schematic for developing the analysis phase of the
compiler for language L using LEX and YACC .
• The analysis phase processes the source program to build an
intermediate representation (IR).
• A single pass compiler can be built using LEX and YACC if the
semantic actions are aimed at generating target code instead of IR.
• Note that the Scanner also generates an intermediate representation
of a source program for use by the parser.
• LEX:

LEX is a computer program that generates lexical

analyzer. LEX accepts an input specification which consist of two
components.

a)The first component is specification of strings representing

the lexical units in L, e g-id’s and constants. The specification is in the
form of regular expressions.

b) The second component is a specification of the semantic

actions aimed at building an IR.

The IR consist of set of tables of lexical units and a

sequence of tokens for the lexical units occurring in a source statement.
Accordingly, the semantic actions make new entries in the tables and
build tokens for the lexical units.

• YACC:
YACC (Yet Another Compiler Compiler) is the standard
parser generator for the UNIX Operating System. Each string
specification in the input to YACC resembles a grammar production.
The parser generated by YACC performs reductions
according to this grammar.The actions associated with a string
specification are executed when a reduction is make according to the
specification.
An attribute is associated with every nonterminal
symbol. The value of this attribute can be manipulated during
parsing .The attribute can be given any user designed structure. A
symbol in the RHS of the string specification ‘$$’represents the
attribute of the LHS symbol of the string specification.
 Unit 2: ASSEMBLERS

 Course points to be covered:

2.1 Elements of assembly language programming
2.2 A simple assembly scheme
2.3 Pass structure of assemblers
2.4 Design of a two pass assembler

 2.1 Elements of assembly language programming:-

Assembly language is a machine dependent & low level
programming language. It provides 3 features given below-

1. Mnemonic operation codes:-

Use of this mnemonic opcodes for machine
instructions eliminates the need to memorize numeric opcodes and it also
enables the assembler to provide helpful diagnostics e.g. indications of
misspelt operation codes.
2. Symbolic operands:-
Symbolic names can be associated with data or
instructions. These names can be used as operands in assembly statements.
The assembler performs memory bindings to these names. This is important
advantage during program modification.
3. Data declarations:-
Data declared in variety of notations including decimal
notations. This avoids manual conversion of constants into their internal
machine representation, for example, conversion of -5 into (11111010)2 or
10.5 into (41A80000)16.

 Statement format:-
An assembly language statement has the following
format:
[Label] <Opcode> <operand spec>[,<operand spec>…]
 where,

[…] indicates enclosed specification is optional.

<operand-spec> has the following syntax:-

<symbolic name> [+<displacement>][(<index register>)]

Thus, some of operand forms are – AREA, AREA+5, AREA(4),

and AREA+5(4).

-1st specification refers- memory word with which the name AREA is
associated.

-2nd specification refers- memory word 5 (displacement/offset from AREA).

-3rd specification refers- indexing with index register 4.

-4th specification- combination of previous two specifications.

 A Simply assembly language:-

In this language , each statement has two

operands, the First operand is always a register which can be any one of
AREG, BREG, CREG and DREG. The Second operand refers to a memory
word using a symbolic name and optional displacement.
 Fig 2.1 Mnemonic Operation Codes

Fig 2.1 lists the mnemonic opcodes for machine instructions. The MOVE
instructions move a value between a memory word and a register. In the
MOVER instruction the second operand is the source operand and first operand
is the target operand. A comparison instruction sets a condition code analogues
to a subtract instruction without affecting the values of operands.

The condition code can be tested by a Branch on Condition (BC)

instruction. The assembly statement corresponding to it has the format

<condition code spec>,< memory address>

It transfers control to the memory word with the address <memory

address> if the current value of condition code matches <condition code spec>.
A BC statement with the condition code spec ANY implies unconditional
transfer of control.

Fig 2.2 Instruction Format

 Fig 2.2 shows the machine instructions format. The opcode, register
operand and memory operand occupy 2, 1 and 3 digits, respectively. The sign is
not a part of instruction. The condition code specified in a BC statement is
encoded into the first operand position using the codes 1-6 for the specifications
LT, LE, EQ, GT, GE and ANY, respectively.

Fig 2.3 An Assembly and Equivalent Machine Language Program

 Assembly Language Statements:-

An assembly language contains three
kinds of statements:-
1. Imperative statements
2. Declaration statements
3. Assembler directives

1. Imperative statements:-
An imperative statement indicates an action to be performed during the
execution of the program. Each imperative statement translates into one
machine instruction.

2. Declaration statements:-
The syntax of declaration statement is as follows:
[Label] DS <constant>
[Label] DC <Value>
The DS is declare storage. The DS statement reserves areas of memory
and associates names with them.
For eg:- The first statement reserves a memory area of 1 word and
associates the name A with it.
The second statement reserves a block of 200 memory words. The name
G is associated with first memory word. Other words can be accessed
through offsets from G.
e.g. :- G+5 for 6th word of memory block etc.
The DC is short for declare constants, and DC statements constructs
memory words containing constants.
The statement:-
ONE DC ‘1’.
associates the name ‘one’ with a memory word containing the value ‘1’.
Constants can be declared by the programmer in different forms-decimal,
binary, hexadecimal.

Use of Constants:-
The DC does not really implement constants, it just
initializes memory words to the given words to the given values. These
values may be changed by moving a new value into the memory word. An
assembly program can use constants in two ways – as immediate
operands and as literals.
The immediate operands can be used in an assembly
statement only if architecture of target machine includes the necessary
features.
Consider the assembly statement
ADD AREG,5
This statement is translated into an introduction with two operands-AREG
and value ‘5’ as an immediate operand.
• A literal is an operand with the syntax=’[value]’
It differs from a constant because its location cannot be specified in the
assembly program. Due to this fact, its value is not changed during the
execution of the program. It differs from an immediate operand because
no specific architecture is needed for its use. An assembler handles a
literal by mapping its use into the features of assembly language.
 3. Assembler Directives:-
Assembler directives instruct the assembler to
perform certain activities during the assembly of a program. Some
assembler directives are:-
a. START [constant]:-
This directive indicates that the first word of the
target program should be placed in the memory word with
address[constant].
b. 2 END [operand spec]:-
This directive indicates the end of the source
program. The [operand spec] indicates the address of the inst.
where execution of the program begin.

 Advantages of Assembly Language:-

The main advantages of assembly language
programming arise from the use of symbolic operand specifications. If there
is some change in program or some statement is inserted in a program, then
in machine language program, this leads to changes in addresses of constants
and memory areas. As a result addresses used in most instructions of the
program had to change. But such changes need not to be made in assembly
language program as operand specifications are symbolic in nature.
 Fig 2.4 Modified assembly and machine language program

Also, assembly language holds an advantage over high

level languages programming in situations where it is necessary to use
specific architectural features of a computer.
e.g. special instructions supported by CPU

 2.2 A Simple Assembly Scheme:-

 Design specification of an assembler:-

To develop a design specification for an assembler a four step approach
is used :-
1. Identify the information necessary to perform a task.
2. Design a suitable data structure to record the information.
3. Determine the processing necessary to obtain & maintain the information.
4. Determine the processing necessary to perform a task.

Here mainly two phases are involved:-

1. Synthesis phase
2. Analysis phase
The fundamental info requirements arise in the synthesis phase. After this , it
is decided that whether this info should be collected during analysis or
derived during analysis.

1. Synthesis phase :-

Consider the assembly statement

MOVER BREG, ONE
To synthesize machine instruction corresponding to this statement, we
must have the following information:-

a. Address of the memory word with which name ‘one’ is associated.

b. Machine operation code corresponding to mnemonic MOVER.

The first item of info. Depends on the source program. So, it must be
made available by analysis phase.
The second item does not depends on source program, it depends on the
assembly language. Hence the synthesis phase can determine this
information itself.

During synthesis phase, two data structures are used:-

1.) Symbol Table
2.) Mnemonics Table
The symbol table’s each entry has two fields:- name and address. This table
is built by analysis phase. An entry in the mnemonics table has two fields:-
mnemonic and opcode.
By using symbol table, synthesis phase obtains the machine address with
which name is associated and by using Mnemonics table, synthesis phase
obtains machine opcode corresponding to a mnemonic.
 2. Analysis phase :-
The primary function performed by the analysis phase
is the building of the symbol table. For this purpose, it must determine the
addresses with which symbolic names used in a program are associated.
Some addresses can be determined directly while some other must be
inferred.
Now, to determine the address of an element, we must
fix the addresses of all elements preceding it. This is called memory
allocation. To implement memory allocation, a data structure called
Location Counter (LC) is introduced. The location counter contains the
address of the next memory word in the target program. LC is initialized
to the constant specified in the START statement.
To update the contents of LC, analysis phase needs to
know the lengths of different instructions. This info. Depends on the
assembly language. To include this info. Mnemonic table can be extended
and a new field called as length is used for this.
The processing involved to maintain the LC is called LC processing.

Fig 2.5 Data Structures of the Assembler

The mnemonic table is a fixed table and merely

accessed by analysis and synthesis phases. The Symbol Table is constructed
during analysis and used during synthesis.
 Tasks performed by Analysis Phases:-
(1) Isolate the label, mnemonic opcode and operand fields of a statement.
(2) If a label is present, enter the pair (symbol) in a new entry of symbol
table.
(3) Check validity of mnemonic opcode by look-up in Mnemonics table.
(4) Perform LC processing i.e. update value contained in LC.

Tasks performed by Synthesis Phase:-

(1) Obtain the machine opcode corresponding to mnemonics.
(2) Obtain address of memory operand from symbol table.
(3) Synthesize the machine form of a constant, if any.

 2.3 Pass Structure of Assemblers:-

The pass of a language processor is one
complete scan of the source program. There are mainly tow assembly schemes:-
(1) Two pass assembly scheme.

(2) One pass assembly scheme

(1) Two Pass Translator / Two Pass Assembler:-

The two pass translator of an assembly language program can handle forward
references easily. The LC processing is performed in the first pass and also the
symbols defined in the program are entered into the symbol table in this pass.
The second pass synthesizes the target form using the address information
found in the symbol table. The first pass performs analysis of the source
program while the second pass performs the synthesis of the target program.
The first pass constructs an intermediate representation (IR) of the source
program for use by the second pass. This representation consists of two main
components – data structures e.g. Symbol table, and a processed form of source
program. The processed form of source program is also called as Intermediate
Code(IC).
 Fig 2.6 Overview of Two pass assembly

(2) Single Pass Translation / Single Pass Assembler:-

In a single pass assembler, the LC processing and construction of symbol
table proceeds as in two pass assembler. The problem of forward references is
tackled using a process called back patching.

Initially, operand field of an instruction containing a forward reference is

left blank. When the definition of forward referenced symbol in encountered, its
address is put into this field which is left blank initially.

Consider the following statement:-

MOVER BREG, ONE

The instruction corresponding to this statement can be only partially

synthesized as ONE is a forward reference. So, the instruction opcode and
address of BREG will be assembled to reside in location 101 and the insertion
of second operand’s address at a later stage can be indicated by adding an entry
of the Table of Incomplete Instructions (TII). This entry is a pair ([instruction
address],[symbol]).

e.g.(101,ONE)

By the time, the END statement is processed, the symbol table would
contain the addresses of all symbols defined in the source program and TII
would contain info. Describing forward references. The assembler can now
process each entry in TII to complete the concerned instruction.

 2.4 Design of Two pass Assemblers:-

Tasks are performed by the passes of two pass assemble are as follows:-

PASS 1:-
(1) Separate the symbol, mnemonic opcode and operand fields.
(2) Build the symbol table.
(3) Perform LC processing.
(4) Construct intermediate representation

PASS 2:-
The Pass 1 performs analysis of the source program and synthesis of the
intermediate representation while Pass 2 processes the intermediate
representation (IR) to synthesize the target program. Before the details of design
of assembler passes, we should know about advanced assembler directives.
 Advanced Assembler Directives:-

(1) ORIGIN:-
The syntax of this directive is
ORIGIN <address spec>
where <address spec> is an <operand spec> or <constant>. This directive
indicates that LC should be set to the address given by <address spec>. The
‘ORIGIN’ statement is useful when the target program does not consist of
consecutive memory words. The ability to use an in the ORIGIN statement
provides the ability to perform LC processing in a relative manner rather
than absolute manner.

(2) EQU:-
<symbol> EQU <address spec>
where <address spec> is an <operand spec> or <constant>. The EQU
statement defines the symbol to represent <address spec>. This differs from
DC/DS statement as associates the name <symbol> with <address spec>.
 Fig 2.7 An assembly program illustrating ORIGIN

(3) LTORG:-
The LTORG statement permits a programmer to specify where
literals should be placed. By default, assembler places the literals after the END
statement. At every LTORG statement, the assembler allocates memory to the
literals of a literal pool. This pool contains all the literals used in the program.
The LTORG directive has less relevance (applicapability) for the simple
assembly languages because allocation of literals at intermediate points in the
program is efficient rather than at the end.

 Pass 1 of the assembler:-

Pass 1 uses the following data structures:-

OPTAB:- A table of mnemonic opcodes and related info.

SYMTAB:- Symbol Table.

LITTAB:- A table of literals used in the program.

OPTAB-contains the fields mnemonic opcode, class and mnemonic
information. The ‘class’ field indicates whether the opcode corresponds to an
imperative statement (IS), a declaration statement (DL) or an assembler
directive (AD). If an imperative statement is present, then the mnemonic info.
Filed contains the pair (machine opcode, instruction length) else it contains the
pair id of a routine to handle the declaration or directive statement.

SYMTAB -contains the field address and length.

LIMTAB -contains the fields literal and address.

 The processing of an assembly statement begins with the processing of its

label field. If it contains a symbol, the symbol and the value in LC is
copied into a new entry of SYMTAB.
 After this, the class field is examined to determine whether the mnemonic
belongs to the class of imperative, declaration or assembler directive
statements.
 If it is an imperative statement, then length of the machine instruction is
simply added to the LC.
 If a declaration or assembler directive statement is present, then the
routine mentioned in the mnemonic info. Field is called to perform
appropriate processing of the statement.
 The LITTAB is used to collect all literals used in the program. The
awareness of different literals pools in maintained by an auxiliary table
POOLTAB. This table contains the literal no. of starting literal of each
literal pool.
 At any stage, the current literal pool is the last pool in LITTAB. On
encountering an LTORG statement (END), literals in the current pool are
allocated addresses starting with current value in LC and LC is
incremented appropriately.
Fig 2.8 Flowchart of Pass 1 Assembler
 Fig 2.9 Data Structures of Pass 1 Assembler

Pass 1 Algorithm: -
(1) Loc ctr = ‘0’ (default value)
Littab ptr = ‘1’
Pool tab ptr = ‘1’

(2) While the next statement is not END statement.

(a) If label is present then
This label = get the name of label
Store [this label, loc counter] in SYMTAB

(b) If an LTORG statement then

(i) Process literals in LITTAB and allocate memory.
(ii) Pooltab ptr = Pool tab ptr + 1;
(iii) Littab ptr = Littab ptr +1;

(c) If instruction is START or ORIGIN then

Loc counter = value specified in operand field;
 (d) If an EQU statement then
(i) this . address = value of [address spec];
(ii) Store [this . label, this . address] in Symbol Table.

(e) If a declaration statement then (DC/DS)

(i) Code = code of declaration statement;
(ii) Size = size of memory area required by DC/DS.
(iii) Loc counter =
(iv) Generate IC (DL, code).

(f) If imperative statement then

(i) Code = machine opcode from OPTAB
(ii) Loc ctr = loc ctr + instruction length from OPTAB
(iii) Generate Intermediate Code (IS, code);
(3) (Processing of END statement)
(a) Perform step 2(b)
(b) Generate IC (AD,02)
(c) Go to Pass 2

Fig 2.10 Error reporting in Pass 1 Assembler

 Intermediate code forms:-

 In this section we consider some variants of intermediate codes and

compare them on the basis of these criteria.
 The intermediate code consists of a set of IC unit, each consisting of the
following three fields.
1. Address
2. Representation of the mnemonic opcode
3. Representation of operands.

Fig 2.11 An IC unit

 Variant forms of intermediate codes, specifically the operand & address

fields, arise in practice due to the trade off between processing efficiency
& memory economy.
 These variants are discussed in separate section dealing with the
representation of imperative statement, & declaration statement &
directive, respectively.
 The information in the mnemonic field is assumed to have the same
representation in all the variants.

Mnemonic field: -
 The mnemonic field contains a pair of the form,
(statement class, code)
where statement class can be IS (Imperative statement), DL(Declaration
statement) or AD (Assembler directives).
 For imperative statement, code is the instruction opcode in the machine
language.
 For declaration and assembler directives, code is an ordinal number within
the class.
 Figure 2.11 shows the codes for various declaration statements and
assembler directives.
 Fig 2.12 Codes for declaration statements and directives

 Intermediate code for imperative statements:-

Variant 1: -
 We consider two variants of intermediate code which differ in the
information contained in their operand fields.
 The address field is assumed to contain identical information in
both variants.
 The first operand is represented by a single digit
• 1-4 for AREG-DREG
• 1-6 for LT-ANY
 The second operand which is a memory operand is represented by
(operand class, code)
where operand class is one of C (constants), S (symbol), L (literals).
• For constant, the code field contains the internal representation of
the constant itself.
• For symbol or literals, code field contain the ordinal number of the
operand’s entry in SYMTAB or LITTAB .
 Fig 2.13 Intermediate code – Variant 1

 Disadvantage: In Variant-I, two kinds of entries may exists in SYMTAB

at any time for defined symbol and for forward references.

Variant 2: -
In variant II,

 For declarative statements and assembler directives, processing of the

operand fields is essential to support LC processing.
 For imperative statements, the operand field is processed only to identify
literal references.
 Symbolic references in the source statement are not processed at all
during Pass-I.

Fig 2.14 Intermediate code – Variant 2

Comparison of the Variants: -
 Variant-1 of the intermediate code appears to require extra work in Pass-
1.
 IC is quite compact in Variant-1.
 Variant-2 reduces the work of pass-1by transferring the burden of
operand processing of pass-1 to pass-2 of the assembler.
 IC is less compact in Variant-2.

Fig 2.15 Memory requirements using

(a) Variant-1
(b) Variant-2
 Processing of declarations and assembler directives:-
 Here we consider the alternative ways of assembler directives.
 DC: A DC statement must be represented in IC. The Mnemonic
field contains the pair (DL, 01). The operand field may contain the
value of the constant in the source form or in the internal machine
representation.
 START or ORIGIN: It is not necessary to retain STRAT and
ORIGIN statement in IC if the IC contains an address field.
 LTORG: Pass I checks for the presence of a literal reference in the
operand field of every statement. If one exists, it enters the literal
in the current literal pool in LITTAB. When an LTORG statement
appears in the source program, it assigns memory addresses to the
literal in the current pool. These addresses are entered in the
address field of their LITTAB entries.
 After performing this fundamental action, two alternatives exist
concerning Pass-I processing.
 Pass I could simply construct an IC unit for the LTORG statement
and leave all subsequent processing to pass II.
 Pass I could itself copy out the literals of the pool into the IC. The
IC for a literal can be made identical to the IC for a DC statement
so that no special processing is required in Pass-II.
 Pass 2 of the assembler:-
Pass 2 Algorithm: -
(1) Code Area Address = address of code area (where target code is to be
enabled)
Pool tab ptr = ‘1’
Loc ctr = o (defined)
(2) While the next statement is not an END statement.
(a) Clear memory buffer area.
(b) If an LTORG statement.
(i) Process the literals and assemble the literals in memory buffer.
(ii) Size = size of memory area reqd. for literals.
(iii) Pooltab ptr = Pool tab ptr + 1

(c) If a START or ORIGIN statement then

 Loc ctr = value specified in operand field.
size = 0

(d) If a declaration statement

(i) If a DC statement
Assemble the constant in memory buffer
(ii) If DS statement
Generate machine code
Size = size of memory req. by DC/DS

(e) If an imperative statement

(i) Assemble instruction in memory buffer.
(ii) Size = size reqd. to store instruction

(f) If size != 0
Store the memory buffer code in code area address.
Loc ctr = loc ctr + size

(3) END statement

(a) Perform steps 2(a) and 2(f)
(b) Write code area into o/p file.
Fig 2.16 Flowchart of Pass-2 Assembler
 Fig 2.17 Error reporting in Pass-2 Assembler

Fig 2.18 Data structures and files in a two pass assembler

 UNIT 3: MACRO AND MACRO
PROCESSORS
 Course points to be covered:
3.1 Macro Definition and Call
3.2 Macro Expansion
3.3 Nested Macro Calls
3.4 Advanced Macro Facilities
3.5 Design of Macro Preprocessor

Macros are used to provide a program generation facility through

expansion. Many languages provide built in facilities for writing macro e.g.
PL/1, C, Ada and C++. When a language does not support built-in macro
facilities then a programmer may achieve an equivalent effect by using
generalized preprocessor or software tools like Awk of UNIX

Macro definition: A macro is a unit of specification for program

generation through expansion.
A macro consists of a name, a set of formal parameters and a body of code.
The use of macro name with a set of actual parameters is replaced by some
code generated from its body. This is called macro expansion.
Difference between macro and subroutines: Use of macro name in the
mnemonic field of an assembly statement leads to its expansion whereas
use of a subroutine name in a call instruction leads to its execution. Thus
the programs using macros and subroutines differ significantly in terms of
program size and execution efficiency.

 3.1 Macro Definition and Call:-

Macro Definition
A macro definition is enclosed between a macro header statement and a macro
end statement. Macro definitions are typically located at the start of a program.
Macro definition consists of:

1. Macro prototype statements

 2. One or more model statements

3. Macro preprocessor statements

1. Macro prototype statements:-

The macro prototype statement declares the name of the macro and the names
and kinds of its parameters. The macro prototype statement has the following
syntax:

<macro name> [<formal parameter specification>[ …..]]

where <macro name> appears in the mnemonic field of assembly Statement and
<formal parameter specification> is of the form

&<parameter name> [<parameter kind>]

2. Model statements:-
A model statement is a statement from which assembly language statements
may be generated during macro expansion.

3. Macro preprocessor statements:-

A preprocessor statement is used to perform auxiliary functions during macro
expansion.

Macro call
A macro is called by writing the macro name in the mnemonic field of an
assembly statement. The macro call has syntax

<macro name> [<actual parameter specification>[,….]]

where actual parameter resembles an operand specification in assembly

statement.

Example 1: The definition of macro INCR is given below

MACRO

INCR &MEM_VAL, &INCR_VAL, &REG

MOVER &REG, &MEM_VAL

 ADD &REG, &INCR-VAL

MOVEM &REG, &MEM_VAL

MEND

The prototype statement indicates that three parameters called MEM_VAL,

INCR_VAL, and REG are parameters. Since parameter kind isn’t specified for
any of the parameters, they are all of the default kind ‘positional parameters’.
Statements with the operation codes MOVER, ADD, MOVEM are model
statements. No preprocessor statements are used in this macro.

 3.2 Macro Expansion:-

A macro leads to macro expansion. During macro expansion, the use of a macro
name with a set of actual parameters i.e. the macro call statement is replaced by
a sequence of assembly code. To differentiate between the original statement of
a program and the statements resulting from macro expansion, each expanded
statement is marked with a ‘+’ preceding its label field.

There are two kinds of expansions

1. Lexical expansion.

2. Semantic expansion

1. Lexical expansion:-
Lexical expansion implies replacement of a character string by another
character string during program generation. It replaces the occurrences of
formal parameter by corresponding actual parameters.

Example 2: Consider the following macro definition

MACRO

INCR &MEM_VAL, &INCR_VAL, &REG

MOVER &REG, &MEM_VAL

 ADD &REG, &INCR-VAL

MOVEM &REG, &MEM_VAL

MEND

Consider macro call statement on INCR

INCR A, B, AREG

The values of formal parameters are:

Formal parameter value

MEM_VAL A

INCR_VAL B

REG AREG

Lexical expansion of the model statement leads to the code

+ MOVER AREG, A

+ ADD AREG, B

+ MOVEM AREG, A

2. Semantic expansion:-
Semantic expansion implies generation of instructions tailored to the
requirements of a specific usage e.g. generation of type specific instruction for
manipulation of byte and word operands. It is characterized by the fact that
different use of macro can lead to codes which differ in the number, sequence
and opcodes of instructions.

Example 3: Consider the macro definition

MACRO

CLEAR &X, &N

LCL &M

&M SET 0

MOVER AREG, =’0’

 .MORE MOVEM AREG, &X+&M

&M SET &M+1

AIF (&M NE N) .MORE

MEND

Consider the macro call CLEAR B, 3

This macro call leads to generation of statements:

+ MOVER AREG, =’0’

+ MOVEM AREG, B

+ MOVEM AREG, B+1

+ MOVEM AREG, B+2

Two key notions of macro expansion are:

Expansion time control flow: This determines the order in which model
statements are visited during macro expansion.

Lexical substitution: It is used to generate an assembly statement from a

model statement.

Flow of control during expansion:

The default flow of control during macro expansion is sequential. In the absence
of preprocessor statements, the model statements of macro are visited
sequentially starting with the statement following the macro prototype statement
and ending with the statement preceding the MEND statement.

Preprocessor statements can alter the flow of control during expansion. So some
model statements are either never visited during expansion or are repeatedly
visited during expansion.

Never visited → conditional expansion.

Repeatedly visited → expansion time loop.

The flow of control during macro expansion is implemented using macro

expansion counter.

Algorithm - Outline of macro expansion

1. MEC: = statement number of first statement following the prototype
statements.

2. While statement pointed by MEC is not a MEND statement

(a)If a model statement then

(i)Expand the statement

(ii)MEC: = MEC+1;

(b) Else (i.e. a preprocessor statement)

(i) MEC: = new value specified in the statement;

3. Exit from macro expansion.

MEC is set to point at the statement following prototype statement. It is

incremented by 1 after expanding a model statement. Execution of a
preprocessor statement can set MEC to a new value to implement conditional
expansion or expansion time loop.

Lexical Substitution
A model statement consists of 3 types of strings:

1. An ordinary string, which stands for itself.

2. The name of formal parameter which is preceded by the character ‘&’

3. The name of a preprocessor variable, which is also preceded by the character
‘&’.

During lexical expansion, strings of type 1 are retained without substitution.

Strings of type 2 and type 3 are replaced by the values of the formal parameters
or preprocessor variables. The rules for determining the value of a formal
parameter depend on the kind of parameter.

Types of parameters
1. Positional parameters.

2. Keyword parameters.

3. Default parameters.

4. Macros with mixed parameter list.

5. Other uses of parameters.

1. Positional parameters
A positional formal parameter is written as &<parameter name>, e.g.
&SAMPLE where SAMPLE is name of parameter. A <actual parameter
specification> is simply an <ordinary string>. The value of positional parameter
XYZ is determined by the rule of positional association as follows:

(i) Find the ordinal position of XYZ in the list of formal parameters in the
macro prototype statement.

(ii) Find the actual parameter specification occupying the same ordinal position
in the list of actual parameters in the macro call statement. Let this be the
ordinary string ABC. Then the value of formal parameter XYZ is ABC.

Example 4: Consider an example

MACRO

INCR &MEM_VAL, &INCR_VAL, &REG

MOVER &REG, &MEM_VAL

ADD &REG, &INCR-VAL

 MOVEM &REG, &MEM_VAL

MEND

In above example, the macro call

INCR A, B, AREG

By rule of positional association, values of the formal parameters are:

Formal parameter value

MEM_VAL A

INCR_VAL B

REG AREG

Lexical expansion of the model statement leads to the code

+ MOVER AREG, A

+ ADD AREG, B

+ MOVEM AREG, A
2. Keyword parameters
A keyword formal parameter is written as ‘&<parameter name>=’.The <actual
parameter specification> is written as <formal parameter name>=<ordinary
string>

The value of a formal parameter XYZ is determined by the rule of keyword

association as follows:

(i) Find the actual parameter specification which has the form XYZ=<ordinary

string>.

(ii) Let <ordinary string> in the specification be the string ABC. Then the value
of formal parameter XYZ is ABC.

The ordinal position of the specification XYZ=ABC in the list of actual

parameters is immaterial. This is very useful in situations where long lists of
used.

Example 5: Consider an example

MACRO

INCR &MEM_VAL=, &INCR_VAL=, &REG=

MOVER &REG, &MEM_VAL

ADD &REG, &INCR-VAL

MOVEM &REG, &MEM_VAL

MEND

In above example, the following macro calls are equivalent.

INCR MEM_VAL=A, INCR_VAL=B, REG=AREG

or

INCR INCR_VAL=B, REG=AREG, MEM_VAL=A

By rule of keyword association, values of the formal parameters are:

 Formal parameter value

MEM_VAL A

INCR_VAL B

Lexical expansion of model statement leads to code.

+ MOVER AREG,A

+ ADD AREG,B

+ MOVEM AREG,A

3. Default specifications of parameters

A default is a standard assumption in the absence of an explicit specification by
the programmer. Default specification of parameters is useful in situations
where a parameter has the same value in most calls. When the desired value is
different from the default value, the desired value can be specified explicitly in
a macro call. This specification overrides the default value of the parameter for
the duration of the call.

The syntax for formal parameter specification is as follows:

&<parameter name>[<parameter kind>[<default value>]]

Example 6: Consider an example
MACRO

INCR &MEM_VAL=, &INCR_VAL=, &REG=

MOVER &REG, &MEM_VAL

ADD &REG, &INCR-VAL

MOVEM &REG, &MEM_VAL

MEND

In above example, the following macro calls are equivalent.

INCR MEM_VAL=A, INCR_VAL=B, REG=BREG

INCR MEM_VAL=A, INCR_VAL=B

INCR MEM_VAL=B, MEM_VAL=A

The first call overrides to us a default specification for the parameter BREG.

By rule of default association, values of the formal parameters are:

Formal parameter value

MEM_VAL A

INCR_VAL B

REG AREG

Lexical expansion of model statement leads to the code.

+ MOVER AREG,A

+ ADD AREG,B

+ MOVEM AREG,A
4. Macro with mixed parameter lists
A macro may be defined to use both positional and keyword parameters. In such
case, all positional parameters must precede all keyword parameters.

Example 7: Consider an example

MACRO

INCR &MEM_VAL=, &INCR_VAL=, &REG=

MOVER &REG, &MEM_VAL

ADD &REG, &INCR-VAL

MOVEM &REG, &MEM_VAL

MEND

In above example, the macro

INCR A, B, REG=BREG

Lexical expansion of model statement leads to the code.

 + MOVER AREG,A

+ ADD AREG,B

+ MOVEM AREG,A

5. Other uses of parameters

The use of parameters is not restricted to operand fields. The formal parameters
can also appear in the label and operand fields of model statements.

Example 8: Consider an example

MACRO

CALC &X, &Y, &OP=MULT, &LAB=

&LAB MOVER AREG, &X

&OP AREG, &Y

MOVEM AREG, &X

MEND
The formal parameters and their values

Formal parameter value

X A

Y B

OP MULT

LAB LOOP

Consider the macro call

CALC A, B, LAB = LOOP

Lexical expansion of model statement leads to the code.

+ LOOP MOVER AREG,A

+ ADD AREG,B

+ MOVEM AREG,A
 3.3 Nested Macro Calls:-
A model statement in a macro may constitute a call on another macro. Such
calls are known as nested macro calls. Expansion of nested macro calls follows
the last-in-first-out (LIFO) rule. Thus, in a structure of nested macro calls,
expansion of the latest macro call (I.e. the innermost macro call in the structure)
is completed first.

Example 9: Consider an example

MACRO

COMPUTE &FRIST, &SECOND

MOVEM BREG, TMP

INCR &FIRST, &SECOND, AREG

MOVER BREG, TMP

MEND

and macro definition

MACRO

INCR &MEM_VAL, &INCR_VAL, &REG=

MOVER &REG, &MEM_VAL

ADD &REG, &INCR-VAL

MOVEM &REG, &MEM_VAL

MEND
 3.4 Advanced Macro Facilities:-
Advanced macro facilities are aimed at supporting semantic expansion. These
facilities can be grouped into

1. Facilities for alteration of flow of control during expansion

2. Expansion time variable

3. Attributes of parameters

1. Alteration of flow of control during expansion:-

Two features are provided to facilitate alteration of flow of control during
expansion:

(i) Expansion time sequencing symbols.

(ii) Expansion time statements AIF, AGO and ANO

Expansion time sequencing symbols

A sequencing symbol (SS) as syntax as:

.<ordinary string>
As SS is defined by putting it in the label field of a statement in the macro body.
It is used as an operand in an AIF or AGO statement to designate the destination
of an expansion time control transfer. It never appears in the expanded form of a
mode statement.

Expansion time statements - AIF, AGO and ANOP

An AIF statement has syntax:

AIF (<expression>) <sequencing symbol>

where <expression> is a relational expression involving ordinary strings, formal

parameters and their attributes, and expansion time variable. If the relational
expression evaluates to true, expansion time control is transferred to the
statement containing <sequencing symbol> in its label field.

An AGO statement has syntax:

AGO<sequencing symbol>

and unconditionally transfers expansion time control to the statement containing

<sequencing symbol> in its label field.

An ANOP statement is written as:

<sequencing symbol>ANOP

and simply defines the sequencing symbol.

Example 10:
MACRO

EVAL &X, &Y, &Z

AIF (&Y EQ &X) .ONLY

MOVER AREG, &X

SUB AREG, &Y

ADD AREG, &Z

AGO .OVER

.ONLY MOVER AREG, &Z

 .OVER MEND

2. Expansion time variable:-

Expansion time variables (EV’s) are variables which can only be used during
the expansion of macro calls. There are two types of EV’s.

(i) Local EV: A local EV is created for use only during a particular macro call
and has following syntax.

LCL <EV specification> [, <EV specification>…]

(ii) Global EV: A global EV exists across all macro calls situated in a program
and can be used in any macro which has a declaration for it. It has following
syntax

GBL <EV specification> [, <EV specification>..]

<EV specification> has a syntax has &<EV name>, where <EV name> is an
ordinary string. Values of EV’s can be manipulated through the preprocessor
statement SET. A SET statement is written as

<EV specification> SET <SET expression>

where <EV specification> appears in the label field and SET in mnemonic field.
A SET statement assigns the value of <SET expression> to the EV specified in
<EV specification>. The value of EV can be used in any field of a model
statement and in the expression of an AIF statement.

Example 11:
LOCAL

MACRO

CONSTANTS

LCL &A

&A SET 1

DB &A

SET &A+1
 DB &A

MEND

A call on macro CONSTANTS is expanded as follow: The local EV A is

created. The first Set statement assigns the value ‘1’ to it. The first DB
statement thus declared a byte constant ‘1’. The second SET statement assigns
the value ‘2’ to A and the second DB statement declares a constant ‘2’.

3. Attributes of formal parameters:-

An attributes is written using the syntax:

<attribute name>’<formal parameter specification>

It represents information about the value of the formal parameter i.e. about
corresponding actual parameter. The type, length and size attributes have the
names T, L and S.

Example 12:
MACRO

DCL_CONST &A

AIF (L’&A EQ 1) .NEXT

.NEXT -----

-----

-----

MEND

Here expansion time control is transferred to the statement having .NEXT in its
label field only if the parameter corresponding to the formal parameter A has
the length of ‘1’.

3.4.1 Conditional expansion:-

Conditional expansion helps in generating assembly code specifically suited to
the parameters in a macro call. This is achieved by ensuring that a model
statement is visited only under specific conditions during the expansion of a
macro. The AIF and AGO statements are used for this purpose.
Example: Consider a macro EVAL such that a call
EVAL A, B, C

generates efficient code to evaluate A-B+C in AREG. When the first two
parameter of a call are identical then EVAL should generate a single MOVER
instruction to load the 3rd parameter into AREG. This is achieved as follows:

Example 13:
MACRO

EVAL &X, &Y, &Z

AIF (&Y EQ &X) .ONLY

MOVER AREG, &X

SUB AREG, &Y

ADD AREG, &Z

AGO .OVER

.ONLY MOVER AREG, &Z

.OVER MEND

Since the value of a formal parameter is simply the corresponding actual

parameter, the AIF statement effectively compares names of the first two actual
parameters. If the names are same, expansion time control is transferred to the
model statement MOVER AREG, &Z. If not, the MOVE-SUB-ADD sequence
is generated and expansion time control is transferred to the statement .OVER
MEND which terminates the expansion.

Expansion time loops

It is often necessary to generate many similar statements during the expansion

of a macro. This can be achieved by writing similar model statement in the
macro.
Example 14: Consider example
MACRO

CLEAR &A

MOVER AREG, =’0’

MOVEM AREG, &A

MOVEM AREG, &A+1

MOVEM AREG, &A+2

MEND

When called as CLEAR B, the statement puts the value 0 in AREG, while the
three MOVEM statements store this value in 3 consecutive bytes with the
address B, B+1, B+2. The same can be achieved by writing an expansion time
loop which visits the model statement, or set of model statements, repeatedly
during macro expansion. Expansion time loops can be written using EV’s and
expansion time control transfer statements AIF and AGO.

Example 15:
MACRO

CLEAR &X, &N

LCL &M

&M SET 0

MOVER AREG, =’0’

.MORE MOVEM AREG, &X+&M

&M SET &M+1

AIF (&M NE N) .MORE

MEND
Consider expansion of macro call

CLEAR B, 3

The macro calls leads to generation of the statements

+ MOVER AREG, =’0’

+ MOVEM AREG, B

+ MOVEM AREG, B+1

+ MOVEM AREG, B+2

3.4.2 Other facilities for expansion time loops:-

The assemblers provide explicit expansion time looping constructs.

1. REPT statement

2. IRP statement

REPT statement
Syntax: REPT<expression>

<expression> should evaluate to a numerical value during macro expansion.

The statements between REPT and ENDM statement would be processed for
expansion <expression> number of times.

Example 16: Following example illustrates the use of this facility to declare
10 constants with the value 1, 2, …, 10

MACRO

CONST10

LCL &M

&M SET 1

REPT 10

DC ‘&M’

&M SETA &M+1

 ENDM

MEND

IRP statement
Syntax: IRP <formal parameter>, <argument list>

The formal parameter mentioned in the statements takes successive values from
the argument list. For each value, the statements between the IRP and ENDM
statements are expanded once.

Example 17:
MACRO

CONSTS &M, &N, &Z

IRP &Z, &M, 7, &N

DC ‘&Z’

ENDM

MEND

A macro call CONSTS 4, 10 leads to declaration of 3 constants with the values

4, 7, 10.

3.4.3 Semantic expansion:-

Semantic expansion is the generation of instructions. It can be achieved by a
combination of advanced macro facilities like AIF, AGO statements and
expansion time variables.

Example 18:
MACRO

CREATE_CONST &X, &Y

AIF (T’ &X EQ B) .BYTE

&Y DW 25

AGO .OVER
 .BYTE ANOP

&Y DB 25

.OVER MEND

This macro creates a constant ‘25’ with the name given by the 2nd parameter
type of the constant matches the type of the first parameter.

Example 19:
MACRO

CLEAR &X, &N

LCL &M

&M SET 0

MOVER AREG, =’0’

.MORE MOVEM AREG, &X+&M

&M SET &M+1

AIF (&M NE N) .MORE

MEND

Example 20:
MACRO

EVAL &X, &Y, &Z

AIF (&Y EQ &X) .ONLY

MOVER AREG, &X

SUB AREG, &Y

ADD AREG, &Z

AGO .OVER

.ONLY MOVER AREG,&Z

.OVER MEND
 3.5 Design of Macro Preprocessor:-
The macro preprocessor accepts an assembly program containing definitions
and calls and translate it into an assembly program which does not contain any
macro definitions or calls. The program form output by the macro preprocessor
can now be handed over to an assembler to obtain the target language form of
the program. Thus the macro preprocessor segregates macro expansion from the
process of program assembly.

The macro preprocessor is economical because it can use an existing assembler.

3.5.1 Design Overview

All tasks involved in macro expansion are

1. Indentify macro calls in a program.

2. Determine the values of formal parameter.

3. Maintain the values of expansion time variables declared in a macro.

4. Organize expansion time control flow.

5. Determine the value of sequencing symbol.

6. Perform expansion of model statement.

The following 4 step procedure is followed to arrive at a design specification for

each task

1. Identify the information necessary to perform a task.

2. Design a suitable data structure to record the information.

 3. Determine the processing necessary to obtain and maintain the
information.

4. Determine the processing necessary to perform the task

Application of this procedure to each of the preprocessor tasks is described in

the following

Identify macro calls in the program

Macro Name Table (MNT) is designed to hold the names of all macro definition
in a program. A macro name is entered in this table when a macro definition is
processed. While processing a statement in the source program, the
preprocessor compares the string found in its mnemonic field with the macro
names in MNT. A match indicate that the current statement is a macro call.

Determine the values of formal parameter.

Actual Parameter Table (APT) is designed to hold the values of formal

parameters during expansion of macro call. Each entry in the table is a pair

(<formal parameter name>, <value>)

Parameter Default Table (PDT) is designed to hold the default values of

keyword parameters during expansion of macro call. Each entry in the PDT is a
pair

(<formal parameter name>, <default value>)

If a macro call statement does not specify a value for some parameter then its
default value would be copied from PDT to APT.

Maintain expansion time variables

Expansion Time Variable’s Table (EVT) is maintained for this purpose. Each
entry in the table is a pair

(<EV name>, <value>)

The value field of a pair is accessed when a preprocessor statement or model

statement under expansion refers to an EV.
Organize expansion time control flow.

Macro definition Table (MDT) is used to store the body of macro. The flow of
control during macro expansion determines when a model statement is to be
visited for expansion.

Determine the values of sequencing symbol.

A sequencing Symbol Table (SST) is maintained to hold this information. The

table contains pair of the form

(<sequencing symbol name>, <MDT entry #>)

where < MDT entry #> is the number of the MDT entry which contains the
model statement defining the sequencing symbol. This entry is made on
encountering a statement which contains the sequencing symbol in its label field
or on encountering a reference prior to its definition.

Perform expansion of a model statement

This is the trivial task given below:

1. MEC points to the MDT entry containing the model statement.

2. Values of formal parameter and EV’s are available in APT and EVT
respectively.

3. The model statement defining a sequencing symbol can be identified

from SST.

Expansion of model statement is achieved by performing a lexical substitution

for the parameters and EV’s used in the model statement.

3.5.2 Data Structure

The tables APT, PDT and EVT contains pairs which are searched using the first
component of the pair as a key e.g. the formal parameter name is used as the
key to obtain its value from APT. This search can be eliminated if the position
of an entity within a table is known when its value is to be accessed.

Thus macro expansion can be made more efficient by storing an intermediate

code for a statement rather than its source form in MDT table.

E.g. MOVER AREG, ABC

Let the pair (ABC, ALPHA) occupy entry#5 in APT. The search in APT can be
avoided if the model statement appears as

MOVER AREG, (P, 5)

in MDT where (P, 5) stands for the words ‘parameter#5’

So the information (<formal parameter name>,<value>) in APT has been split

into two tables:

PNTAB contains formal parameter names

APTAB contains formal parameter values

Similarly analysis leads to splitting of EVT into EVNTAB and EVTAN and
SST into SSNTAB and SSTAB. PDT is replaced by a keyword parameter
default table (KPDTAB).

Table of macro preprocessor

 PNTAB and KPDTAB are constructed by processing the prototype
statement.
 Entries are added to EVNTAB and SSNTAB as EV declaration and SS
definition/references are encountered.
 MDT entries are constructed while processing model statement and
preprocessor statement in macro body.
 An entry is added to SSTAB when the definition of a sequencing symbol is
encountered.
 APTAB is constructed while processing a macro call.
 EVTAB is constructed at the start of expansion of macro.
Example 21:
MACRO

CLEAR_MEM &X, &N, &REG=AREG

LCL &M

&M SET 0

MOVER &REG, =’0’

.MORE MOVEM &REG, &X+&M

&M SET &M+1

AIF (&M NE N) .MORE

MEND

Consider the macro call

CLEAR_MEM AREA, 10, AREG

ALGORITHM FOR PROCESSING OF A MACRO DEFINITION

1. SSNTAB_PTR := 1;

PNTAB_PTR := 1;

KPDTAB_PTR := 1;

SSTAB_PTR := 1;
 MDT_PTR := 1;

2. Process the macro prototype statement and form the MNT entry

a) name := macro_name;

b) For each positional parameter

i) Enter parameter name in PNTAB[PNTAB_PTR]

ii) PNTAB_PTR := PNTAB_PTR+1;

iii) #PP := #PP+1;

c) KPDTP := KPDTAB_PTR;

d) For each keyword parameter

i) Enter parameter name and default value in KPDTP[KPDTP_PTR]

ii) Enter parameter name in PNTAB[PNTAB_PTR]

iii) KPTAB_PTR := KPDTP_PTR+1

iv) PNTAB_PTR := PNTAB_PTR+1

v) #KP := #KP+1

e) MDTP := MDT_PTR

f) #EV := 0

g) SSTP := SSTAB_PTR

3. While not a MEND

a) If LCL statement then

i. Enter expansion time variable name in EVNTAB

ii. #EV := #EV+1

b) If a model statement then

i. If a label field contain a sequencing symbol then

If symbol is present in SSNTAB then

 q := entry number in SSNTAB

else

enter symbol in SSNTAB[SSNTAB_PTR]

q := SSNTAB_PTR

SSNTAB_PTR := SSNTAB_PTR+1

SSTAB[SSTP+q-1] := MDT_PTR

ii. For a parameter, generate the specification (P, #n)

iii. For an expansion variable, generate the specification (E , #m)

iv. Record the IC in MDT[MDT_PTR]

v. MDT_PTR :=MT_PTR+1

4. MEND statement

If SSNTAB+PTR=1 then

SSTP=0

Else

SSTAB_PTR = SSTAB_PTR + SSNTAB_PTR-1

If #KP: = 0 then KPCTP: = 0

ALGORITHM FOR MACRO EXPANSION

1. Perform initializations for the expansion of a macro

a) MEC := MDTP field of the MNT entry

b) Create EVTAB with #EV entries and set EVTAB-PTR

c) Create APTEB with #PP+#KP entries and set APTAB_PTR

d) Copt keyword parameter defaults from the entries

KPDTAB[KPDTP]……KPDTAB[KPDTP+#KP-1] into
 APTAB[#PP+1]…..APTAB{#PP+#KP}

e) Process positional parameters in the actual parameter list copy them

into APTAB{1]>>>>>APTAB[#PP

f) For keyword parameters in the actual parameter list

Search the keyword name in parameter name field of

KPDTAB[KPDTP]….KPDTAB[KPDTP+#KP-1].Let KPTDAB[q]
contain a matching entry. Enter value of the keyword parameter in the
call in APTAB[#PP+q-KPDTP+1].

2. While statement pointed by MEC is not MEND statement

a) If model statement then

i. Replace the operands of the (P, #n) and (E , #m) by values in

APTAB[n] and EVTAB[m] respectively.

ii. Output the general statement.

iii. MEC := MEC+1

b) If a SET statement with the specification (E, #m) in the label field then

i. Evaluate the expression in the operand and set an appropriate

value in EVTAB[m].

ii. MEC:= MEC+1

c) In an AGO statement with (S, #s) in operand field then

MEC := SSTAB[SSTP+s-1];

d) If an AIF statement with (S, #s) in operand field then

If condition in the AIF statement is true then

MEC := SSTAB[SSTP+s-1]

3. Exit from macro expansion.

3.5.3 DESIGN OF A MACRO ASSEMBLER:-
The use of macro preprocessor followed by a conventional assembler is an
expensive way of handling macros. Since the number of passes of source
program is large and many functions get duplicated. For example, analysis of
source statement to detect macro calls requires us to process the mnemonic
field. A Similar function is required in the first pass of the assembler. Similar
functions of preprocessor and the assembler can be merged if macros are
handled by a macro assembler which performs macro expansion and program
assembly simultaneously. This also reduces the number of passes.

Pass structure of a macro assembler

To design the pass structure of a macro assembler we identity the

functions of a macro preprocessor and the conventional assembler which can be
merged. After merging, the functions can be structured into passes of the macro
assembler.

Pass I

1. Macro definition processing.

2. SYMTAB construction.

Pass II

1. Macro expansion.

2. Memory allocation and LC processing.

3. Processing of literals.

4. Intermediate code generation.

Pass III

1. Target code generation.

The pass structure can be simplified if attributes of actual parameters are not to
be supported. The macro preprocessor would then be a single pass program.
Integrating pass I of the assembler with preprocessor would give us the
following two pass structure
Pass I

1. Macro definition processing.

2. Macro expansion.

3. Memory allocation, LC processing and SYMTAB construction.

4. Processing of literals.

5. Intermediate code generation.

Pass II

1. Target code generation.

 Unit 4: Compiler and Interpreters

Course points to be covered:

4.1 Aspects of Compilation

4.2 Static and Dynamic Memory Allocation
4.3 Compilation of Expression
4.4 Compilation of Control Structures
4.5 Interpreters

4.1 Aspects of Compilation:-

 Two aspects:
– Generate code.
– Provide Diagnostics.
 To understand the implementation issue, we should know PL
features contributing to semantic gap between PL and Execution domain.
 PL Features:
1. Data Types
 2. Data Structure
3. Scope Rules
4. Control Structures

1. Data type
 Definition: A data type is the specification of
(i) Values those entities of the type may have
(ii) Operations that performed on entities of type
 The following tasks are involved:
1. Check legality of operation for types of operand
2. Use type conversion operation
3. Use appropriate instruction sequence of the target machine
var
x, y : float;
i,j : integer;
Begin
y := 10;
x := y + i;
Type conversion of i is needed.
i: integer;

CONV_R AREG, I
ADD_R AREG, Y
MOVEM AREG, X

2. Data Structure
• PL permits declaration of DS and use it.
• To compile the reference of element of DS compiler must develop memory
mapping to access allocated area.
• A record, heterogeneous DS leads to complex memory mapping.
• User defined DS requires mapping of different kind.
• Proper combination of DS is required to manage such complexity of structure.
• Two kind of mapping is involved
– Mapping array reference
– Access field of record

Example:
Program example (input, output);
type
employee = record
name : array [1...100] of character;
sex : character;
id: integer
end;
var
info : array [1..500] of employee;
i,j : integer;
begin {main program}
info[i].id := j;
end
3. Scope Rules
• Determine the accessibility of variable declared in different blocks of a program.
• E.g.: x, y : real;
y, z : integer;
x := y; B A
• Stat x:=y uses value of block ‘B’.
• To determine accessibility of variable compiler performs operation:
– Scope Analysis
– Name Resolution

4. Control Structure
• Def: It is collection of language features for altering flow of control during
execution.
• This includes:
– Conditional transfer of control
– Conditional execution
– Iterative control
– Procedure call
• Compiler must ensure non-violation of program semantics.
Eg: for i := 1 to 100 do
begin
lab1 : if i = 10 then…
end;
• Forbidden: control is transferred to label1 from outside the loop.
• Assignment statements are also not allowed.
4.2 Memory Allocation:-

•Memory Allocation involves Three Important task:

– Determine memory requirement to represent value of data items.
– Determine memory allocation to implement lifetime & scope of data item.
– Determine memory mapping to access value in non-scalar data items.
• Topic list:

A. Static and dynamic memory allocation

B. Memory allocation in block structured language

C. Array allocation & access

[A] Static & Dynamic Allocation

Static Memory Allocation:
1. Memory is allocated to variable before execution of program begins.

2. Performed during compilation.

3. At execution no allocation & de-allocation is performed.

4. Allocation to variable exists even if program unit is not active.

5. Variable remains allocated permanently till execution does not end.

6. E.g.: Fortran
7. No. of flavors or types

8. Adv: Simplicity and faster access.

9. Memory wastage compared to dynamic.

Fig. Static memory allocation.

Dynamic Memory Allocation:

1. Memory allocation of variable is done all the time of execution of program.

2. Performed during execution.

3. Allocation & de-allocation actions occurs only at execution.

4. Allocation to variable exists only if program unit is active.

5. Variables swap from and to allocation state & free state.

6. E.g.: PL/I, ADA, Pascal.

7. Two types/ flavors

- 1. Automatic allocation
 - 2. Program controlled allocation

8. Adv: Recursion support DS size dynamically.

9. Less memory wastage compared to formal.

Fig. Dynamic memory allocation

[B] Memory allocation in Block Structured Language

• Block contain data declaration

• There may be nested structure of block

• Block structure uses dynamic memory allocation

• E.g.: PL/I, Pascal, ADA

• Sub Topic List

1. Scope Rules

2. Memory Allocation & Access

3. Accessing Non-local variables

 4. Symbol table requirement

5. Recursion

6. Limitation of stack based memory allocation

1. Scope Rules:
• If variables var (i) is created with name name(i) in block B.

• Rule 1: var (i) can be accessed in any statement situated in block B.

• Rule 2: Rule 1 + ‘B’ is enclosed in ‘B’ unless ‘B’ contains declaration using
same name name(i)

• E.g.: variables B = local

B’ = non-local

A{

x,y,z : integer;

B{ g : real;

C{ h,z : real;

}C

}B

D{ i,j : integer;

}D

}A
 fig.Block Structured Program.

2. Memory allocation & Access

• Implemented using extended stack model.
• Automatic dynamic allocation is implemented using extended stack model.
• Minor variation: each record has two reserved pointers, determining its scope.
• AR: Activation Record
• ARB: Activation Record Block
• See the following figure:
– When execution starts, state changes from (a) to state (b).
– When execution exit block C, state change (a) to (b).
• Allocation:(Actions at block entry)

1. TOS := TOS + 1;

2. TOS* := ARB;

3. ARB := TOS;

4. TOS := TOS + 1;

5. TOS* := ...........(sec reserved pointer 2)

6. TOS := TOS + n;

• So address of ‘z’ is <ARB> + z;

 De-allocation:

1. TOS := ARB – 1;

2. ARB := ARB* ;
3. Accessing non-local variables
– n1_var : is a non local variable

– b def : block defined into

– b use : block in use

• Then textual ancestor of block b use is a block which encloses block b use that
is b def.

• Level m ancestor is a block which immediately encloses level m-1 ancestor.

• S nest b use: static nesting level of block b use.

• Rule: when b use is in execution, b def must be active.

• i.e.ARBb _def exist in stack while execution ARb_use.

• n1 _var are accessed by start address of ARb_def + dn1_var where, dn1_var is

displacement of n1_var in ARb_def.

• Lets learn few more terms

– Static Pointers

– Display

(i) Static Pointer:

• Access of non-local variables is implemented using second reserved pointer in

AR.

• It has 1 (ARB) 0 (ARB)

• 1 (ARB) is static pointer.

• At the time of creation of AR for Block B its static pointer is set to point AR of
static ancestor of b.

• Access non-local variables:

1. r := ARB;

2. Repeat step-3 m times

3. r := 1(r)

4. Access n1_var using address <r> + dn1_var.

• Example: Status after execution:

– TOS := TOS + 1

– TOS* := address of AR at level 1 ancestor.

– <r> to access x at statement x:= z

• r:= ARB;

• r:= 1(r);

• r:= 1(r);

• Access x using address <r> + dx.

 Fig. Accessing non-local variables with static pointers.

(ii) Displays:

• Display (1) = address of level (S_nest(b)-1) ancestor of B.

• Display (2) = address of level (S_nest(b)-2) ancestor of B.

• Display [S_nestb-1] = address of level 1 ancestor of B.

• Display [S_nestb-1] = address of ARB.

• For large value of level difference, it is expensive to access non- local variables
using static pointers.

• Display is an array used to improve the efficiency of non-local variables

accessibility.
• Eg: r:= display[1] access x using address <r> + dx.

fig.Accessing non-local variables with displays.

4. Symbol Table Requirement:

• To improve dynamic allocation & access, compiler should perform following
task.

– Determine static nesting level of b_current

– Determine variable designated with scope rules.

– Determine static nesting level of block & dv.

– Generate code.

• Extended stack model is used because it has

 – Nesting level of b_current

– Symbol table for b_current.

• Symbol table has

– Symbol

– Displacement.

fig.Stack Structured Symbol Table.

5. Recursion:
• Extended stack model best for recursion.

6. Limitations of stack based memory allocation:

• Not good for program controlled memory allocation.
• Not adequate for multi-active programs.

[C] Array allocation & Access

• A[5,10]. 2D array arranged column wise.
• Lower bound is 1.
• Address of element A[s1,s2] is determined by formula:
• Ad.A[s1,s2] = Ad.A[1,1] + { (s2-1)xn + (s1-1)} x k.
• Where,
– n is number of rows.
– k is size, number of words required by each element.
• General 2D array can be represented by a[l1:u1,l2:u2]
• The formula becomes
Ad.A[S1,S2] = Ad.A[l1,l2] + {(s2-l2) x (u1-l1+1) + (s1-l1)}xk.
• Defining the range:
– Range1: u1-l1+1
– Range2: u2-l2+1
• Ad.A[s1,s2]
= Ad.A[l1,l2] + {(s2-l2) x range1 + (s1-l1)} x k.
= Ad.A[l1,l2] – (l2 x range1 + l1) x k + (s2 x range1+s1) xk.
= Ad.A[0,0] + (s1 x range1 + s1) x k.
• We can compute following values at compilation time or allocation time
(given that subscripts are constants).
fig.Memory allocation for a 2D array
Dope Vector:

• Definition: Dope vector is an descriptor vector

– Accommodated
commodated in symbol table

– Used by generation of code phase.

• No. of dimension of array determine format & size of DV.

• Code generation use d.Dv to check validity of subscripts.

4.3 Compilation of Expression:-
A. A Toy generator for expression
a)Operand Descriptor
b)Register Descriptor
c)Generating an Instruction
d)Saving Partial Results

B. Intermediate code for expression

a)Postfix string
b)Triples & Quadruples
c)Expression Trees

[A] A Toy Generator for Expressions

•Major issues in code generation for expression:
–Determination of execution order for operators.
–Selection of instruction used in target code.
–Use of registers and handling partial results.
a) Operand Descriptor:

•Attributes:
•Addressability:
•Specifies:
–Where operand is located
–How it can be accessed.
•Addressability Code:
–M : operand is in memory
–R : operand is in register
–AR : address is in register
–AM : address is in memory
•Address:
–Address of CPU register or memory

•Operand descriptor is build for every operand that is id’s, constant, partial results.
–PRi:-Partial Results
–Opj:-Some operator
•Operand descriptor is an array in which operand descriptions are stored.
•Descriptor # is a descriptor in operand descriptor array.
b) Register Descriptor
•It has 2 fields:
–Status: free / occupied
–Operand Descriptor #
•Stored in register descriptor array.

•One register descriptor exist for each CPU register.

•Register Descriptor for AREG, after generating code for a*b.

c) Generating on Instruction
•Rule: any one operand need to be in register to perform any operation over it.

•If not, one of the operand is brought to register.

•Function code genis called with OP; and descriptors of its operands as parameters.
d) Saving Partial Results
•If all the registers are occupied, register are freed by transferring content of
temporary location in memory.
•r is available to evaluate operator OPi.
•Thus, ‘temp’ array is declared in target program to hold partial results.
•Descriptor of partial result must change/modify when partial result are moved to
temporary location.
•After partial result a*b is moved to temp location.

[B] Intermediate Codes for Expressions

a) Postfix Strings
b) Triples and Quadruples
c) Expression Trees

a) Postfix Strings
•Here each operand appear immediately after its last operand.
•Thus, operators can be evaluated in order in which they appear in string.
•Eg: a+b*c+d*e^fabc*+def^*+
•We perform code generation from postfix string using stack of operand
descriptors.
•Operand appears and then operand descriptors are pushed to the stack i.e. stack
would contain descriptor fro a,b & c when first * is encountered.
•Little modification in extended stack model can manage postfix string efficiently.

b) Triples & Quadruples

1) Triples:
•Triple is a representation of elementary operations in the form of a pseudo
machine instruction
•Slight change in algorithm of operator precedence help us convert infix string to
triples.
Example: Triples for a+b*c+d*e^f
Operator Operand 1 Operand 2
* B C
+ 1 A
^ E F
* D 3
+ 2 4

2) Indirect Triples:
•Are useful in optimizing compiler.

•This arrangement is useful to detect the identical expression occurrence in the

program.

•For efficiency Hash Organization can be used for the table of triples.

•Indirect triple representation provide memory economy.

•Aid: Common sub expression elimination.

•See figure 6.20 and example 6.22 on pg.no.189.

3) Quadruple:
•Result name: designates result of evaluation that can be used as operand for other
quadruple.
•More convenient than using triples & indirect triples.
•For example : a+b*c+d*e^f
•Remember, they are not temporary locations (t) but result name.
•For elimination, result name can become temporary location.
c) Expression Tree
•Operator are evaluated in order determined by bottom up parsing which is not
most efficient.
•Hence, compiler’s back-end analyze expression to find best evaluation order.
•What will help us here?
•Expression Tree: is a AST (Abstract Syntax Tree) which depicts the structure of
an expression.
•Thus, simplifying analysis of an expression to determine best evaluation order.
•How to determine best evaluation order for the expression?
–Step 1: Register Requirement Label (RR Label) indicates no. of register required
by CPU to evaluate sub-tree.
–Step 2: Top Down parsing and RR Label information is used and order of
evaluation is determined.

4.4 Compilation of Control Structure:-

Topic List:
A. Control Transfer, Conditional Execution & Iterative Constructs
B. Function & Procedure Calls
C. Calling Conventions
D. Parameter Passing Mechanism
a) Call by Value
b) Call by Value Result
c) Call by Reference
d) Call by Name

Definition: Control Structure

•Control structure of a programming language is the collection of language features
which govern the sequencing of control through a program.

•Control structure of PL construct for

–Control Transfer
–Conditional Execution
–Iterative Construct
–Procedure Call
[A] Control Transfer, Conditional Execution & Iterative Construct
•Control transfer is implemented through conditional & un-conditional ‘goto’
statements are the most primitive control structure.
•Control structure like if, for or while cause semantic gap b/w PL domain &
execution domain.
•Why? Control transfer are implicit rather then explicit.
•This semantic gap is bridge in two steps:
–Control structure is mapped into goto.
–Then this programs are translated to assembly program.

[B] Function & Procedure Call

x := fn-1(y , z) + b * c;
•In the given statement function call on fn-1 is made after execution returns the
value to calling function.
•This may result to some side effects.
•Def: Side Effect:Side effect of a function (procedure) call is a chance in the value
of a variable which is not local to the called function (procedure).
•A procedure call only achieves side effect, it doesn’t return a value.
•A function call achieves side effect also returns same value.
•Compiler must ensure following for function call:
–1. Actual (y,z) parameters are accessible in called function.
–2. Called function is able to produce side effect according to rules of PL.
–3. Control is transferred to, and is returned from the called function.
–4. Function value is returned to calling program.
–5. All other aspects of execution of calling program are unaffected by function
call.

•Compiler uses set of features to implement functions:

–Parameter List: contains descriptor for each parameter.
–Save Area: Register Save Area
–Calling Conventions: there are few execution time assumptions:
i. How parameter list is accessed?
Ii. How save area is accessed?
 iii. How transfer of control at call & return are implemented?
iv. How function value is returned to calling program.
–(iii) & (iv) are performed by machine instruction or CPU
–Others are assumptions they are not achieved by CPU or machine instruction.

[D] Parameter Passing Mechanism

•Define semantics of parameter usage inside a function.
•Thus, defines the kind of side effects, a function can produce on its actual
parameter.
•Types:
–Call by value
–Call by value result
–Call by reference
–Call by name

1. Call by Value
•Actual parameters are passed to called function.
•These values are assigned to corresponding formal parameters.
•Values are passed in ‘one direction’. i.efrom calling program to called program.
•If function changes value of formal parameter, changes are not reflected on actual
parameter.
•Thus, can’t produce any side effect on parameters.
•Generally used in built in function.
•Advantage:
–Simplicity
–Efficient if parameters are scalar variables.

2. Call by Value Result

•Extends capability of call by value.

•Copies values of formal parameter back to corresponding actual parameter at

return.

•So, side effects are reflected at return.

•Advantage:
–Simplicity
•Disadvantage:
–Incurs higher overhead.

3. Call by Reference
•Address of actual parameter is passed to called function.
•Parameter list is actually the list of addresses.
•At every access, corresponding actual parameter is obtained from parameter list.
•Code:
1. r <-<ARB> + (dDp)AR orr <-<r_par_list> + (dDp)par_list
2. Access value using address contained in register.
•Code analysis:
Step 1: incurs overhead
Step 2: produces instantaneous side effects.
•Mechanism is popular because has clear semantics.
•Plays important role at the time of nesting of structures.
•How? It provides updated value.
•See eg. 6.29 on pg. no. 197.
•Here, z,iare non local variables of alpha.
•Alpha be called as (d[i],x).
•Value of ‘x’ changes as ‘b’ also changes.

4. Call by Name
•Same effect as call by reference.
•Every occurrence of formal parameters in the called function is replaced by the
name of the corresponding actual parameter.
•Eg:

a = d[i];
z = d[i];
i= i+ 1;
-b = d[i] + 5;
x = d[i] + 5;
•Achieves instantaneous side effects.
•Has implication of changes in parameters during execution.
•Code:
1. r <-<ARB> + (dDp)AR
r <-<r_par_list> + (dDp)par_list
2. Call the function whose address is contained in r.
3. Use the address returned by the function to access p.
•Advantage:
Changes are made dynamically which makes call by name mechanism
immensely powerful.
•Dis-advantage:
High overhead at step 2.
Not much practically practiced.
4.5 INTERPRETERS:-
 Both Compilers and Interpreters analyze a source statement to determine its
meaning.
 During compilation, analysis of statement is followed by code generation
and during interpretation, it is followed by actions which implements its
meaning.

Compiler v/s Interpreter

• Notation:

– Tc : Average Compilation time per statement

– Te : Average Execution time per statement

– Ti : Average Interpretation time per statement

• Here we assure Tc = Ti.Te

• Size(p) = number of statements in program P.

• Stmt_executed (p) = number of statements executed in Program P.

Example:

• Let,

– Size = 200

– 20 statements are executed

– Loop has 10 iterations

– Loop has 8 instructions in it

– 20 stmts are followed by loop for printing result.

• Then, stmt_executed (p) = 20 + (10 * 8) +20= 120.

• Total Execution Time:

For compilation model: 200. Tc + 120*.Te= 206Tc

For interpretation model: 120.Tc

• Conclusion: Clearly interpretation is better than compilation as far as execution

time is concerned.

Why Use Interpreter?

• Efficiency & Certain environmental benefits.

• If required modification, recommended when stmt execution is less than size P.

• Preferred during program development.

• Also when programs are not executed frequently

• Best for writing programs for an editor, user interface, OS development.

Components:

1. Symbol Table: Holds information concerning entities present in program.

2. Data Store: Contains values of data items declared.

3. Data Manipulation Routines: A set containing a routine for every legal data
manipulation actions in the source language.

• Advantages:

– Meaning of source statement is implemented using interpretation routine which

results to simplified implementation.

– Avoid generation of machine language instruction.

– Helps make interpretation portable.

– Interpreter itself is coded in high level language.

A Toy Interpreter:

 Design and operation of an interpreter for basic return in Pascal

 The data store of interpreter consist of two large array

-rvar to store real

-ivar to store integer

 Last few locations in the rvar and ivar arrays are used as stacks for
expression evaluations with the help of the pointers r_tos and i_tos
respectively.
Pure & Impure Interpreters:

 Pure Interpreter:
-In a pure interpreter, the source program is retained in the source form all
through its interpretation.
-This arrangement incures substantial analysis overheads while interpreting
a statement .

Data

Source
Interpreter Result
program

 Impure Interpreter:

-It performs some preliminary processing of source program to reduce

analysis overhead during interpretation.

-The preprocessor converts the program to an intermediate representation .

-E.g. Intermediate code(IC) can be analyzed more efficiently than the source
form of the program.

Data

Source
Preprocessor Interpreter Result
program
 Unit 5: Linkers

Course points to be covered:

5.1 Relocation and Linking Concepts
5.2 Design of a Linker
5.3 Self-Relocating Programs
5.4 Linking for Overlays
5.5 Loaders

Execution of Program written in a Language L involves the following

steps:
1.Translation of a Program
2.Linking of Program with other Programs needed for its Execution
3.Relocation of the Program to execute from the specific memory area
allocated to it.
4.Loading of Program in the memory for the execution.
These steps perform by:
Step 1 is performed by translator for languages L, Step 2 & 3 are
performed by Linker and Step 4 is performed by a loader.

Figure 5.1 contains the schematic showing steps 1-4 in the execution of
the program.The translator outputs a program from calledanobjectmodule
for the program. The linker processes a set of object modules to produce a
ready-to-execute program form, which we will call a binaryprogram. The
loader loads this program into the memory for the purpose of execution.
As shown in the schematic, the object module(s) and ready to execute
program form can be stored in the form of files for repeated use.

Translated, linked and load time address:

While compiling the program P, translator is given an origin specification

for P. this is called a translated origin of P. the translator uses the value of
translated origin to perform memory allocations for the symbols declared
in P. this result in the assignment of a translation time addresstsymbto each
symbol symb in the program. The execution start address or simply the
start address of a program is the address of the instruction from which its
 execution must begin. The start address specified by the translator is the
translated start address of the program.

translator linker loader Binaryprogram

Control flow

result
Source
Program object module binary program
Fig.5.1 A schematic representation of program execution.
The origin of a program may have to be changed by the linker or loader
for one of the two reasons. First, the same set of translated addresses may
have been used in different object modules constituting a program, e.g.
object modules of library rotines often have the same translated origin.
Memory allocation to such programs would conflict unless their origins
are changed. Second, an operating system may require that a program
should execute from a specific area of memory. This may require a
change in its origin. The change of origin leads to changes in the
execution start address and in the addresses assigned to symbols. The
following terminology used to refer to the address of a program entity at
different times:
1. Translation time (or translated) address: Address assigned by the
translator
2. Linked address: Address assigned by the linker
3. Load rime (or load) address: Address assigned by the loader
The same prefixes translation time (or translated), linked and load time
(or load) are used with the origin and execution start address of a
program. Thus,
The same prefixes translation time (or translated) , linked and load time
(or load) are used with the origin and execution start address of a
program. Thus ,
1.Translated origin: Address of the origin assumed by the translator This
I is the address specified by the programmer in an ORIGN statement.
2.Linked Origin: Address of the origin assigned by the linker while
producing a binary program.
3.Load origin: Address of the origin assigned by the loader while loading
the program for execution.
The linked and load origins may differ from the translated origin of a
program due to one of the reasons mentioned earlier.

Example 5.1 Consider the assembly program and its generated code
shown in Fig. 5.2 The translated origin of the program is 500. The
translation tine address of LOOP is there fore 501. If the program is
loaded for execution in the memory area starting with address 900.the
load time origin is 900. The load time address of LOOP would be 901.

Statement Address Code

START 500
THE TRANSLATED
ENTRY TOTAL
ORIGIN
EXTERN MAX,ALPHA
READ A 500) +90 0 540
LOOP 501)
. THE TRANSLATED TIME ADDRESS
OF LOOP
.
MOVER AREG,ALPHA 518) +04 1 000
BC ANY,MAX 519) +06 6 000
.
.
BC LT,LOOP 538) +06 1 501
STOP 539) +00 0 000
A DS 1 540)
TOTAL DS 1 541)
END

Fig. 5.2.A sample assembly program and its generated code.

5.1 Relocation and Linking Concepts:-
5.1.1 Program Relocation:-

Let AA be the set of absolute addresses-instruction or data addresses-used

in theinstructions of a program P. AA ≠ φ implies that program P
assumes its instructions and data to occupy memory words with specific
addresses. Such a program-called an address sensitive program- contains
one or more of the following:
1. An address sensitive instruction: an instruction which uses an address
aiꞒ AA
2. An address constant: a data word which contains an address aiꞒAA

In the following, we discuss relocation of programs containing address

sensitive instructions. Address constants are handled analogously.
An address sensitive program P can execute correctly only if the
start address of the memory area allocated to it is the same as its
translated origin. To execute correctly from any other memory area, the
address used in each address sensitive instruction of P must be corrected.

Definition 5.1 (Program relocation): Program relocation is the process

of modifying the addresses used in the address sensitive instructions of a
program such that the program can execute correctly from the designated
area of memory.
If linked origin ≠ translated origin, relocation must be performed
by the linker. If load origin ≠ linked origin, relocation must be performed
by the loader. In general, a linker always performs relocation, whereas
some loaders do not. For simplicity, in the first part of the chapter it has
been assumed that loaders do not perform relocation-- that is, load origin
= linked Origin. Such loaders are called absolute orders. Hence the terms
'load origin' and linked origin are used interchangeably. However, it
would have been more precise to use the term linked origin. (Loaders that
perform relocation, i.e. relocating loaders, are discussed in Section 7.6)

Example 5.2 The translated origin of the program in Fig.5.2 is 500. The
translation time address of symbol A is 540. The instruction
corresponding to the statement READ A (existing in translated memory
word 500) uses the address 540, hence it is an address sensitive
instruction. If the linked origin is 900, A would have the link time address
940. Hence the address in the READ instruction should be corrected to
940. Similarly, the instruction in translated memory word 538 contains
501, the address of LOOP. This should be corrected to 901. (Note that the
operand addresses in the instructions with the addresses 518 and 519 also
need to be corrected. This is explained in Section 5.1.2.)

Performing relocation:

Let the translated and linked origins of program P be t_origin and

l_origin, respectively. Consider a symbol symb in P. Let its translation
time address be tsymb and link time address be lsymb. The relocation factor
of P is defined as
o relocation_factorP = l_originP – t_originP
(5.1)
Noe that relocation_factor can be positive, negative or zero
Consider a statement which uses symb as an operand. The translator puts
the
address tsymb in the instruction generated for it. Now
o tsymb = t_originP + dsymb

wheredsymb is the offet of symb in P. Hence

o lsymb = l_originP+ dsymb

Using 5.1

o lsymb = toriginP + dsymb+ relocation_factorP

 = toriginP + dsymb+ relocation_factorP
lsymb = tsymb + relocation_factorP (5.2)

Let IRRP designate the set of instructions requiring relocation in program

P. Following (5.2), relocation of program P can be performed by
computing the relocation factor for P and adding to the translation time
address(es) in instruction i ꞒIRRP
.
Example 5.3 For the program of Fig. 5.2
relocation factor = 900 - 500
= 400.
Relocation is performed as follows: IRRP contains the instructions with
translated dresses 500 and 538, The instruction with translated address
300 contains the address 540 in the operand field. This address is changed
to (540 + 400)=940. Similarly, 400 is added to operand address in the
instructions with translated address 538. This achieves relocation
explained in Ex. 5.2

5.1.2 Linking:-

Consider an application program AP consisting of a set of program units

SP={Pi}A program unit Pi interacts with another program unit Pj , by
using the address of the Pj’s instructions and data in its own instructions.
To realize such interactions must contain public definitions and external
as references as defined in the following:

Public definition: a symbol pub_symb defined in a program unit

whichmay be referenced in other program unit
External reference:a reference to a symbol ext_symb which is not
defined in the program unit containing the reference

the handling of public definitions and external references is described in

the following.

EXTRN and ENTRY statements:

The ENTRY statement lists the public definitions of a program unit,i.e

lists those symbols defined in the program unit which may be referenced
in other program units.The EXTRN statement lists the symbols to which
external references are made in the program unit.

Example 5.4 In the assembly program of Fig. 5.2, the ENTRY statement
indicates that a public definition of TOTAL exists in the program. Note
that LOOP and A are not public definitions even though they am defined
it the program. The EXTRN statement indicates that the program contains
external references to MAX and AL PHA. The assembler does not know
the address of an external symbol. Hence it puts zeros in the address
fields of the instructions corresponding to the statements MOVER
AREG, ALPHA and ВС ANY. MAX. If the EXTERN statement did not
exist, the assembler would have flagged references to MAX and ALPHA
as errors.
Resolving external references:

Before the application Program AP can be executed, it is necessary that

for each Pi in SP, every external reference in Pi should be bound to the
correct link time address.

Definition 5.2 (Linking) : Linking is the process of binding an external

reference to the correct link time address. An external reference is said to
be unresolved until linking is performed for it is said to be resolved when
its linking is completed.

STATEMENT ADDRESS CODE

START 200
ENTRY ADRESS
--
ALPHA DS 25 231) +00 0 025
END

Fig. 5.3 Program unit Q

Example 5.5. Let the program unit of Fig 5.2 (referred to as program unit
P) be linked with the program unit Q described in Fig. 5.3
Program unit P contains an external reference to symbol ALPHA which
is a public definition in Q with the translation time address 231. Let the
link origin of P be 900 and its size be 42 words. The link origin of Q is
therefore 942, and the link time putting the link time address of ALPHA
is 973. Linking is performed by putting the link time address of ALPHA
in the instruction of P using ALPHA, i.e. by putting the address 973 in
the w instruction with the translation time address 518 in P.
Binary Programs:

Definition 5.3 (Binary Programs)

A Binary Programs is a machine language program comprising of a set of
Program units SP such that ꓯ Pi Ꞓ SP.

1. Pi has been relocated to the memory area starting at its link origin, and
2. Linking has been performed for each external reference Pi.

To form a binary program from a set of object modules, the programmer

invokes the linker using the command
linker<link origin>, <object module names> [,<execution start
address>]
where<link origin> specifies the memory address to be given to the first
word of the binary program <execution start address> is usually a pair
(program unit name, offset in program unit). The linker converts this into
the linked start address. This is stored along with the binary program for
use when the program is to be executed. If of <execution start address> is
omitted the execution start address is assumed to be the same as the
linked origin.

5.1.3 Object Modules:-

The object module of a program contains all information necessary to

relocate and link the program with other programs.
The object module of a program P Consists of 4 components :

1. Header: The header contains translated origin, size and execution start
address of P.
2.Program: This component contains the machine language program
corresponds to P.
3.Relocation table: (RELOCTAB) This table describes IRRp. Each
RELOCTAB entry contains a single field:
Translation address : translated address of an address sensitive
instructions
4.Linking table (LINKTAB): This table contains information concerning
thepublic definition and external references in P.
 Each LINKTAB entry contains three fields:

Symbol: Symbolic name

Type:PD/EXT indicating whether public definition or external reference
Translated address: For a public definition, this is the address of the first
memory word allocated to the symbol. For an external reference, it is the
address of the memory word which is required to contain.
the address of the symbol.

Example 5.6 Consider the assembly program of Fig.5.2. The object

module of the program contains following information
1.translated origin = 500, size = 42, execution start address = 500
2.Machine language instructions shown in fig 5.2
3.Relocation table
500
538

4.Linking table

ALPHA EXT 518

MAX EXT 519

A PD 540

5.2 DESIGN OF A LINKER:-

5.2.1 Relocation and Linking Requirements In Segment
Addressing:-

Relocation and Linking Requirements In Segmented Addressing. The

relocation requirements of a program are influenced by the addressing
structure of the computer system on which it is to execute. Use of the
segmented addressing structure reduces the relocation requirements of a
program)
Example 5.7. Consider the program of Fig 5.4 Written in the assembly
language of Intel 8088. The ASSUME statement declares the segment
registers CS and DS to be available for memory addressing. Hence all
memory addressing is performed by using suitable displacements from
their contents. Translation time address of A is 0196. In statement 16, a
reference to A is assembled as a displacement of 196 from the contents of
the CS register. This avoids the use of an absolute address, hence the
instruction is not address sensitive Now no relocation is needed if
segment SAMPLE is to be loaded in the memory, starting at the address
2000 because the CS register would be loaded with the address 2000 by a
calling program (or by the OS) The effective operand address would be
calculated as <CS> +0196, which is the correct address 2196. A similar
situation exists with the reference to B in statement 17, The reference to
B is assembled as a displacement of 0002 from the contents of the DS
register. Since the DS register would be loaded with the execution time
address of DATA_HERE, the reference to B would be automatically
relocated to the correct addresses.

Sr. Statement Offset

No.
0001 DATA_HERE SEGMENT
0002 ABC DW 25 0000
0003 B DW? 0002
. .
. .
0012 SAMPLE SEGMENT
0013 SEGMENT CS:SAMPLE,
DS:DATA_HERE
0014 MOV AX,DATA_HERE 0000
0015 MOV DS,AX 0003
0016 JMP A 0005
0017 MOV AL,B 0008
 .
0027 A MOV AX,BX 0196
.
0043 SAMPLE ENDS
0044 END

Fig 5.4 An 8088 assembly program for linking

Though use of segment registers reduces the relocation requirements, it

does notcompletely eliminate the need for relocation. Consider statement
14 of Fig. 5.4, viz. MOV AX, DATA_HERE which loads the segment
base of DATA_HERE into the AX register preparatory to its transfer into
DS register. Since the assembler knows DATA_HERE to be a segment it
makes provision to load the higher order 16 bits of the address of
DATA_HERE into the AX register.It does not know the link time address
of the DATA_HERE, hence it assembles the MOV instruction in the
immediate operand format and puts zeros in the operand field. It also
makes the entry of this instruction in RELOCTAB so that linker will put
the appropriate address in the operand field. Inter-segment calls and
jumps handled in similar ways.
Relocation is some what more involved in some of intra-segments jumps
assembled in the FAR FORMAT. For example, consider following
program:

FAR_LAB EQU THIS FAR : FAR_LAB is a FAR label

JMP FAR_LAB : A FAR jump

Here the displacement and the segment base of FAR_LAB are to be put
in the JMP instruction itself. The assembler puts the displacement of
FAR_LAB in the first two operand bytes of the instruction, and makes a
RELOCTAB entry for the third and fourth operand bytes which are to
hold the segment base address. A statement like
ADDR_A DW OFFSET A
(which is an ‘address constant') does not need any relocation since the
assembler can itself put the required offset in the bytes.

5.2.2. Relocation Algorithm:-

Algorithm 5.1 (Program Relocation)

1.program_linked_origin = <link origin> from linker command;

2.For each object module

(a)l_origin = translated origi of the object module;

OM_size = size of the object module;
(b)relocation_factor = program_linked_origin – t_origin;
(c)Read the machine language program in work area
(d)Read RELOCTAB of the object module
(e) For each entry in RELOCTAB
(i)translated_addr = address in the RELOCTAB entry;
(ii)address_in_work_area = address of work_area +
tranalated_address – t_origin;
(iii) Add relocation_factor to the operand address in the
word with the address address_in_work_area.
(f) program_linked_origin:= program_linked_origin + OM_size;

The computations performed in the algorithm are along the lines

described in Section 5.1.1The only new action is the computation of the
work area address of theword requiring relocation (step 2(e)(üi). Step 2(f)
increments program_linked_originso that the next object module would
be granted the next available load address.

Example 5.8 Let the address of work_area be 300. While relocating the
object module of Ex. 5.6, relocation factor=400. For the first
RELOCTAB entry, address_in_work_area= 300+500-500= 300, This
word contains the instruction for READ A It is relocated by adding 400
t0 the operand address in it. For the second RELOCTAB entry, address in
work area 300 +538-500= 338. The instruction in this word similarly
relocated by adding 400 to the operand address in it.

5.2.3 Linking Requirements:-

Features of a programming language influence the linking In Fortran all

program units are translated separately. Hence all subprogram Common
variable references require linking, Pascal procedures are typically nested
inside the main program. Hence procedure references do not require
linking they can be handled through relocation. References to built in
functions, however, require linking. In C program files are translated
separately. Thus, only function calls cross file boundaries and references
to global data require linking.
A reference to an external symbol alpha can be resolved only if
alpha is de-clared as a public definition in some object module. This
observation forms the basis of program linking. The linker processes all
object modules being linked and builds a table of all public definitions
and their load time addresses. Linking for alpha is simply a matter of
searching for alpha in this table and copying its linked address into the
word containing the external reference.
A name table (NTAB) is defined for use in program linking. Each
entry of the table contains the following fields:

Symbol : symbolic name of an external reference or an object module

Linked_address: Fora public definition, this field contains linked ad-dress

of the symbol. For an object module, it con-tains the linked origin of the
object module.

Most information in NTAB is derived from LINKTAB entries with type

= PD.

Algorithm 5.2(Program Linking)

1.program_linked_origin = <link origin> from linker command;
2.For each object module
(a)t_origin = translated origin of the object module;
OM_size = size of the object module;
(b)relocation_factor = program_linked_origin – t_origin;

(c)Read the machine language program in work_area

(d)Read RLINKTAB of the object module
(e)For each entry in LINKTAB with type=PD
name= symbol;
linked_address = tranalated_address + relocation_factor;
Enter (name, linked_address) in NTAB;
(f)Enter (object module name, program_linked_origin) in NTAB.
(g)program_linked_origin = program_linked_origin + OM_size;
3.For each object module

(a)t_origin = = translated origin of the object module;

program_linked_origin = load_address from NTAB;
(b) For each LINKTAB entry with type = EXT
(i)address_in_work_area: = address of work-area
+program_linked_origin - <link_origin>+translated_address
– t_origin
(ii)Search symbol in NTAB and copy its linked address. Add
the linkedaddress to the operand address in the word with the
address address_in_work_area.

Example 5.9: While linking program P of Ex.5.2 and program of Ex.5.5

with linked origin 900, NTAB contains the following information

Symbol Linked
address
P 900
A 940
Q 942
ALPHA 973

Let the address of work_area be 300. When the LINKTAB entry of

ALPHA is processed during linking, address_in_work_area=300+900-
900+518 500, i.e. 318 Hence, the linked address of ALPHA, ie. 973, is
copied from the NTAB entry of ALPHA and added to the word in
address 318 (step 3(b)(ii)).

Exercise 5.2
1.It is required to merge a set of object modules {omi} to construct a
single object module om’. This reduces the linking and relocation time in
situations where the object modules in {omi} are independent- that is,
they call or use each other
(a)What are the public definitions of ‘om’ ?
(b)What are the external references in om’?
(c)Explain how the RELOCTAB’s can be merge.

2.Comment on feasibility of reconstructing the original object modules

from om’. Is any additional information required to facilitate it?
5.3 SELF-RELOCATING PROGRAMS:-
The manner in which a program can be modified, or can modify itself, to
execute from a given load origin can be used to classify programs
into the following:
1. Non-relocatable programs
2. Relocatable programs
3. Self-relocating programs.

A non-relocatable program is a program which cannot be executed in any

memory area other than area starting on its translated origin. Non-
relocability is the address sensitive of program and lack of information
concerning the address sensitive instructions in the program. The
difference between the Relocatable and Non-relocatable is availability of
information concerning the address sensitive instructions in it. A
relocatable program can be processed it relocate it to desired area of
memory.
A self-relocating program is a program which-can perform the
relocation of its own address
sensitive instructions. It contains the following two provisions for this
purpose:
1.A table of information concerning the address sensitive
instructions exists as apart of the program.
2. Code to perform the relocation of address sensitive instructions
also exists asa part of the program. This is called the relocating logic.

The start address of the relocating logic is specified as the

execution start address of the program. Thus the relocating logic gains
control when the program is loaded in memory for execution. It uses the
load address and the information concerning now transferred to the
relocated program.
A self-relocating program can execute in any area of the memory.
This is very important in time sharing operating systems where the load
address of a program is likely to be different for different executions.

Exercise 5.3
1. Comment on following statements:
(a) Self-relocating programs are less efficient than relocatable programs.
 (b) There would be no need for linkers if all programs are coded as self-
relocating programs.
2.A self-relocating program needs to find its load address before it can
execute its relocating logic. Comment on how this information can be
determined by the program.

5.4 LINKING FOR OVERLAYS:-

Definition 5.4 (Overlays): An overlay is a part of program (or software
package) which has the same load origin or some other part(s() of the
program.
Overlays are used to reduce the main memory requirement of the
program.

Overlay Structured Program:

We refer to program containing overlays as an overlays structured

program. Such a program consists of
1.A permanently resident portion, called the root
2.A set of overlays
Execution of an overlay structured program proceeds as follows: To start
with, the root is loaded in memory and given control for the purpose of
execution. Other over lays are loaded as and when needed. Note that the
loading of an overlay overwrites a previously loaded overlay with the
same load origin. This reduces the memory requirement of a program. It
also makes it possible to execute programs whose size exceeds the
amount of memory which can be allocated to them. The overlay structure
of a program is designed by identifying mutually exclusive modules-that
is, modules which do not call each other. Such modules do not need to
reside simultaneously in memory. Hence they are located in different
overlays with the same load origin.

Example 5.15: Consider a program with 6 sections named init, read,

trans_a, trans_b, trans_c, and print. Init performs some initializations and
passes control to read. Read reads one set of data and invokes one of
trans_a, trans_b ortrans_c depending on the values of the data, print is
called to print the results. Trans_a, trans_b and trans_c are mutually
exclusive. Hence, they can be made into I overlays. Read and print are
put in the root of the program since they are needed for each set of data.
For simplicity, we put init also in the root, though it could be made into
an overlay by itself. Figure 5.9 shows the proposed structure of the
program. The overlay structured program can execute in 40 K bytes
though it as a total size of 65 K bytes. It is possible to overlay parts of
trans_a against each other by analyzing its logic. This will further reduce
the memory requirements of the program.The overlay structure of an
object program is specified in the linker command.
o 0
 init init
10K 10K
 read read

15K 15K
 trans_a write
35K
 trans_b 20K
50K 30K
trans_c 35K trans_c
60K trans_b
write 40K
65K trans_a

Fig. 5.9 An Overlay tree

Example 5.16. The MS-DOS LINK command to implement the overlay

statement of fig. 5.9 is

LINK init +read + write + (trans_a) + (trans_b) + +(trans_c), <execution

file>,<library files>

The object module(s) within parenthesis become one overlay of the

program. The object module not included in any overlay become part of
the root. LINK produces a single bimary Program containing all overlays
and stores it in <executable file>.

Example 5.17 the IBM mainframe linker commands for the overlay
structure of Fig.5.9 are as follows:
Phase main: PHASE MAIN, +10000
INCLUDE INIT
INCLUDE READ
INCLUDE READ
Phase trans_a: PHASE A+TRANS, *
INCLUDE TRANS_A
Phase trans_b: PHASE B_TRANS, A_TRANS
INCLUDE TRANS_B
Phase trans_c: PHASE C_TRANS, A_TRANS
INCLUDE TRANS_C

Each overlay forms a binary program (i.e. a phase) by itself. A PHASE

statement specifies the object program name and linked origin for one
overlay. The linked origin can be specified in many different ways. It can
be absolute address specification as in the first PHASE statement. In
second PHASE statement, the ‘*’ indicates contents of the counter
program_linked_origin. Thus, the linked origin is the next available
memory byte. The specification in the last two PHASE statements
indicate the same linked origin as that of A_TRANS.

Execution of an overlay structured program:

For linking and execution of an overlay structured program in MS DOS

the linker produces a single executable file at the output., which contains
two provisions to support overlays. First, an overlay manager module is
included in the executable file.
This module is responsible for loading the overlays when needed.
Second, all calls at cross overlay boundaries are replaced by an interrupt
producing instruction. To start with, the overlay manager receives control
and loads the root. A procedure call which crosses overlay boundaries
leads to an interrupt. This interrupt is processed by the overlay manager
and the appropriate overlay is loaded into memory.
When each overlay is structured into a separate binary program as in IBM
main- frame systems a call which crosses overlay boundaries leads to an
interrupt which is attended by the OS kernel. Control is now transferred
to the OS loader to load the appropriate binary program.

Changes in LINKER algorithms:

The basic change required in LINKER algorithms of Section7.4.1 is in

the assignment of load addresses to segment program_load_origin can be
used as before while processing the root portion of a program. This size
of root would decide the load address of the overlays, program
load_origin would be initialized to this value while processing every
overlay. Another change in the LINKER algorithm would be in the
 handling of procedure calls that cross overlay boundaries. LINKER has
toidentify an inter-overlay call and determine the destination overlay.
This information must be encoded in the software interrupt instruction.
An open issue in the linking of overlay structured programs is
the handling ofobject modules added during auto-linking. Should these
object modules be added to the current overlay or to the root of the
program? The former has the advantage of cleaner semantics (think of
procedures with static or own data), however it may increase the memory
requirement of the program.

Exercise 5.5

1. A call matrix (CM) is a square matrix with boolean values in which

CM[I,j]=true only if procedure I calls procedure j (directly or
indirectly). Describe tow a programmer can use this information to
design the overlay structure of a program.

2. Modify the passes of LINKER to perform the linking of an overlay

structured program. (Hint: Should each overlay have a separate
NTAB?)

3. This chapter has discussed static relocation. Dynamic relocation

implies relocation during the execution of a program, e.g. a program
executing in one area of memory may be suspended, relocated to
execute in another area of memory and resumed there.
Discuss how dynamic relocation can be performed.

4. An inter-overlay call in an overlay structured program is expensive

from the execution viewpoint. It is therefore proposed that if sufficient
memory is available during program execution, all overlays can be
loaded into memory s01iat each inter-overlay call can simply be,
executed as an inter-segment call. Comment on the feasibility of This
proposal and develop a design for it.

5.5 LOADERS:-
As described in Section 5.1.1 an absolute loader can only load programs
with load origin linked origin. This can be inconvenient if the load
address of a program is likely to be different for different executions of a
program. A relocating loader performs relocation while loading a
program for execution. This permits a program to be executed in different
parts of the memory.

Linking and loading in MS DOS:

MS DOS operating system supports two object program forms.

A file with a.COM extension contains a non-relocatable object program
whereas a file with a.EXE extension contains a relocatable program. MS
DOS contains a program EXE2BIN which converts a .EXE program into
a COM program. The system contains two loaders-an absolute loader and
a relocating loader. When the user types a filename with the .COM
extension, the absolute loader is invoked to simply load the object
program into the memory. When a filename with the .EXE extension is
given, the relocating loader relocates the program to the designated load
area before passing control to it for execution.

Exercise 5.6

1. Compare the .COM files in MS-DOS with an object module. (MS-

DOS object modules are contained in files with .OBJ extension).
2. It is proposed to performed dynamic linking and loading of the
program units constituting an application. Comment on the nature of
the information which would have to be maintained during the
execution of a program.
 UNIT 6: SOFTWARE TOOLS
 Course points to be covered:
6.1 Software Tools for Program Development
6.2 Editors
6.3 Debug Monitors
6.4 User Interfaces

Computing involves two activities-

1)program development.
2)use of application.

Definition: A software tool is a system program which

1)interfaces a program with the entity generating its input data. OR
2)interfaces the results of a program with the entity consuming them.
The entity generating the data or consuming the results may be a
program or a user.

Fig 6.1 Schematic of a Software Tool

 6.1 Software Tools for Program Development

The fundamental steps in program development are:
1. Program design ,coding and documentation
 2. Preparation of programs in machine reachable form
3. Program translation ,linking and loading
4. Program testing and debugging
5. Performance tuning
6. Reformatting the data and/or results of a program to suit other
programs.

1. PROGRAM DESIGN AND CODING:-

Two categories of tools used in program design and coding are

i. program generators
ii. programming environments.
Program Generator generates a program which performs a set of
functions described in its specification. Use of a program generator saves
substantial design effort since a programmer merely specifies what functions a
programs should perform rather than how the functions should be implemented.
Coding effort is saved since program is generated rather than coded by hand.

A Program Environment supports coding by incorporating

awareness of the programming languages syntax and semantics in the language
editor.

2. PROGRAM ENTRY AND EDITING:-

These tools are text editors as front ends.

Two modes
1)command mode
2)data mode.
In COMMAND MODE, it accepts user commands specifying the
editing function to be performed.

In DATA MODE, the user keys in the text to be added to file.

Failure to recognize the current mode of the editors can lead to mix up of
commands and data. This can be avoided in two ways- ONE approach, a quick
exit is provided from data mode. ANOTHER popular approach is to use the
screen mode, wherein the editor is in the data mode most of time.
3. PROGRAM TESTING AND DEBUGGING:-

 Important steps in testing and debugging are selection of test data for the
program.
 Software tools to assist the programmer in these following steps:
1. Test data generators help the user in selecting test data for his
program. Their use helps in ensuring that a program is thoroughly
tested.
2. Automated test drivers help in regression testing .
3. Debug monitor help in obtaining information for localization of errors.
4. Source code control system help to keep track of modification in the
source code.
Test data selection uses the notion of an execution path which is
sequence of program statements visited during an execution. For testing, a
program can be viewed as a set of execution paths. A test data is a set of input
values which satisfy these conditions.

4. ENHANCEMENT OF PROGRAM PERFORMANCE:-

 Program efficiency depends on two factors:

1)the efficiency of the algorithm.
2)the efficiency of its coding.
 Program designer improve its efficiency of an algorithm by rewriting it.
 This is time consuming process.[RO parts& it is less than three percent of
program code].
 A profile monitor is a software tool that collects the information regarding
the execution behaviour of a program.
 Using this information the programmer can focus attention on the program
section consuming a significant amount of execution time.

5. DESIGN OF SOFTWARE TOOLS:-

Program Processing and Instrumentation
 Program processing techniques used to support static analysis of programs .
 Program instrumentation implies insertion of statement in the program.
 During insertion ,the inserted statement perform a set of desired function.
 Whereas in debug monitors an inserted statement indicate that execution has
reached a specific point in a source program.
 The instrumented program is translated using a standard translator.

Program Interpretation and Program generation

Fig 6.2 Software tools using interpretation and program generation

 Software tools using the technique of interpretation and program

generation.
 An interpreter based tools is only as portable as the interpreter it uses.
 A generated program is more efficient and portable.
 However, interpreter based tools suffer from poor efficiency and poor
portability, since an interpreter based tool is only as portable as the
interpreter it uses.
 6.2 Editors
Text editors come in the following forms:
1. Line editors.
2. Steam editors.
3. Screen editors.
4. Word processor.
5. Structure editors.

 The scope of edit operation in a line editor is limited to a line of text.

 Line and steam editors typically maintain multiple representation of text .
 The primary advantage of line editors is their simplicity.
 A stream editor views the entire text as a stream of characters.
 This permits edit operations to cross line boundaries.
 Stream editors typically support character, line context oriented commands
based on current editing context indicated by position of a text pointer.
 The pointer can be manipulated using positioning or search commands.

6.2.1 Screen editors:-

 A line or stream editor does not display the text in manner it would appear
if printed.
 A screen editor uses what you see is what you get principle in editor design.
 The editor displays a screenful of a text at a time.
 The user can move the cursor over the screen, position it at a point where
he desires to perform some editing and proceed with editing directly.
 Thus it is possible to see the effect of an edit operation on screen.
 This is very useful while formatting the text to produce printed documents.

6.2.2 Word Processors:-

 Word processors are basically document editors with additional features to
produce well formatted hard copy outputs.
 Essential features of word processors are commands for moving sections of
text from one place to another, merging of text, and searching and
replacement of words.
 Many word processors support a spell-check option.
 Wordstar is a popular editor of this class.

6.2.3 Structure Editors:-

 A structure editors incorporates an awareness of the structure of a
document.
 The vi editor of UNIX and the editor in desk top publishing system are
example of these.
 This is useful in browsing through a document.
 A special class of structure editors , called syntax directed editors, are used
in programming environments.

6.2.3 Design of an Editors:-

 The fundamental function in editing are travelling, editing, viewing and
display .
 Travelling implies movement of the editing context a new position within
the text.
 Viewing implies formatting the text in a manner desired by the user.
 The display component maps the views into the physical character of the
display device being used.
 The separation of viewing and display function give rise to interesting
possibilities like multiple window ,on the same screen ,concurrent editing
operations using the same display terminals etc.
 The editing and viewing filters operate on the internal form of text to
prepare the forms suitable of editing and viewing.
 These forms are put in the editing and viewing buffers respectively.
 The viewing and display manager makes provision for appropriate display
of this text.
 Apart from the fundamental editing functions, most editors support an undo
function to nullify one or more of the previous edit operations performed by
the user.
 Multilevel undo commands pose obvious difficulties in implementing
overlapping edits.
 Fig 6.3 Editor structure

 6.3 Debug Monitors:-

 Debug monitor provide the following facilities for dynamic debugging.

1. Setting breakpoint in the program .

2. Initiating a debug conversation when control reaches a breakpoint.
3. Displaying value of variables.
4. Assigning new values to variables.
5. Testing user defined assertion and predicates involving program
variables.

 The debug monitor functions can be easily implemented in an interpreter.

 The sequence of steps involved in dynamic debugging of a program is as
follows:
 1. The user compiles the program under the debug option. The compiler
produces two files – the compiled code file and the debug information
file.
2. The user activates the debug monitor and indicates the name of
program to be debugged. The debug monitor opens the compiled code
and debug information files for the program.
3. The user specifies his debug requirements- a list of breakpoints and
actions to be performed at breakpoints. The debug monitor instruments
the program, and builds a debug table containing the pairs.
4. The instrumented program gets control and executes up to breakpoint.
5. A software interrupt is generated when the <Sl_instrn> is executed.
Control is given to the debug monitor which consults the debug table
and performs the debug actions specified for the breakpoint.
6. Steps 4 and 5 are repeated until the end of the debug session.
UNIX supports two debuggers- sdb which is a PL level debugger and adb which
is an assembly language debugger, debug of IBM PC is an object code level
debugger.

6.3.1 Testing assertions:-

 A debug assertions is a relation between the values of program variables.
 An assertions can be associated with a program statement.
 The debug monitor verifies the assertion when execution reaches that
statement .
 Program execution continues if the assertion is fulfilled , else a debug
conversation is opened.
 Use of debug assertions eliminates the need to produce voluminous
information for debugging purposes.

PROGRAMMING ENVIROMENTS
 A programming environment is a software system that provides integrated
facilities for program creation, editing, testing and debugging .it consist of
the following component.
1. A syntax directed editor(which is structure editor)
2. A language processor –a compiler ,interpreter, or both
3. A debug monitor.
4. A dialog monitor.
 All components accessed through the dialog monitor.
 The syntax directed editor incorporates a front end for the programming
language.
 Editor performs syntax analysis and converts it into an intermediate
representation(IR).
 The compiler and the debug monitor share the IR.
 If a compiler is used ,it is activated after the editor has converted a
statement to IR.
 At any time during execution the programmer can interrupt program
execution and enter the debug mode or return the editor.
 The system may provide the program development and testing function.

 6.4 User Interfaces:-

 A user interface (UI) plays a vital role in simplifying the interaction of a
user with an application.
 UI functionalities have two important aspects-issuing of command and
exchange of data.
 In those days UI’s did not have an independent identity.
 The on line help an application component of the UI serves the function of
educating the user in the capabilities of an application and modalities of its
usage.
 A well designed UI make the difference between a gruding set of users and
an excited user population for an application.
 A UI can be visualised to consist of two components-
1)a dialog manager
2)a presentation manager

6.4.1 Command Dialogs:-

 Commands are issued to an application through a command dialogs are:
1)command languages
2)command menus
3)direct manipulation
1. Command Languages:-
 Command languages for computer application are similar to command
languages for operating system.
 Syntax is <action><parameter>.
 More sophisticated command languages have both declarative and
imperative commands and a syntax and semantics of their own.
 A practical difficulty in the use of the command language is the need to
learn it before using the application.
 This implies a large commitment of time and effort on part of the user,
which makes casual use of the application forbidding.
2. Command Menus:-
 Command menus provide obvious advantages to the casual user of an
application .
 A hierarchy of menus can be used to guide the user into the details
concerning a functionality.
 Interesting variations like pull-down menus are designed to simplify the use
of menus systems.
 Most turbo compilers use pull down command menus.

3. Direct Manipulation:-
 A direct manipulation system provides the user with a visual display of the
universe of the application.
 The display shows the important objects in the universe.
 Actions or operations over objects are indicated using some kind of pointing
devices.
 Eg. A cursor or a mouse.

PRINCIPLES OF COMMAND DIALOG DESIGN

1. Ease of use
2. Consistency in command structure
3. Immediate feedback on user commands
4. On line help to avoid memorizing command details
5. Undo facility
6. Shortcuts for experienced users.
6.4.2 Presentation of Data:-
 Data for an application can be input through free from typing.
 Application results can be presented in the form of tables.
 Summary data can be presented in the form of graph, pie charts etc.

6.4.3 On Line Help:-

 On line help is very important to promote and sustain interest in the use of
an application.
 It minimizes the efforts of initial learning of commands and avoids the
distraction of having to consults a printed document to resolve every doubt.
 On line help to the users can be organized in the form of an on line
explanation .
 Another effective method of organizing on line help is to provide context
sensitive help, whereby a query can fetch different responses depending on
the current position of the user in the application.

Hypertext:-
 Hypertext visualizes a document to consist of a hierarchical arrangement of
information units and provide a variety of means to locate the required
information. this takes the form of
1. Tables and indexes
2. String searching function
3. Means to navigate within the document
4. Backtracking facilities
 6.4.4 Structure of a User Interface :-

 Fig shows UI schematic using a standard graphics package .

 The UI consist of two main components, a presentation manager and a
dialog manager.
 The presentation manager is responsible for managing the user’s screen and
for accepting data and presenting results.
 The dialog manager is responsible for interpreting user commands and
implementing them by invoking different modules of an application code.
 The dialog manager is also responsible for error message and on line help
functions and for organizing change in the visual context of the user.

6.4.5 User Interface Management Systems :-

 A user interface management system (UIMS) automates the generation of
user interfaces.
 The (UIMS) accepts specification of the presentation and dialog semantics to
produce the presentation and dialog managers of the UI respectively.
 The presentation and data manager could be generated programs specific to
the presentation and dialog semantics.
 A variety of formalisms have used to describe dialog semantics.
 In a grammar based description, the syntax and semantics of commands are
specified in a YACC like manner.
 A screen with icons is displayed to the user.
 The grammar and event description approaches lack the notion of a sequence
of actions.
 The finite state machine approach can efficiently incorporate this notion.
 The basic principle in this approach is to associate a finite state machine with
each window or each icon.

MENULAY
 Menulay is an early UIMS using the screen layout as the basis for the dialog
model the UI designer starts by designing the user screen to consist the set of
icons.
 This action is performed when icon is selected .
 The interface consists of set of screens the system generate set of icon tables
giving the name and description of an icon and a list when an event is
selected.

HYPERCARD
 This UIMS from apple incorporates object orientation in the event oriented
approach .
 A card has an associated screen layout containing buttons and fields .
 A button can be selected by clicking the mouse on it .
 A field contain editable text.
 Many cards can share the same background.
 A hypercard program is thus a hierarchy of card called a stack.
 Action foreign event is determined by using hierarchy of cards as an
inheritance hierarchy.
 Hypercard uses an interpretive schematic to implement UI.