Programming Languages

1
Programming Languages

• "A computer program does what you tell it to do,
  not what you want it to do."    
•  Greer's Third Law.

2
Programming Language Toolset
C
If all you have is a hammer, then everything
looks like a nail.
3
More than just a tool

• Current view from carpentry
• A hammer is more than just a hammer. It's a
  personal tool that you get used to and you form a
  loyalty with. It becomes an extension of
  yourself."
• http//www.hammernet.com/romance.htm

4
Objectives

• Explore Programming Language history and design
• Review Programming Paradigms
• Learn about the Traditional Programming concepts
• Review Procedural Units
• Understand the basics of Program Translation
• Reading Brookshear Ch. 5

5
Language Generations
Five Generations Procedural Languages 1GL
Machine language 2GL Assembly language 3GL
High-level language Nonprocedural Languages 4GL
Very high-level language 5GL Natural language

• Low levels closer to binary format
• High levels - closer to human language

6
Machine Language

• Written in strings of 0 and 1
• Only language the computer understands
• All other programming languages are translated to
  machine language
• Computer dependent
• Recall that CISC instructions wont work for a
  RISC architecture

7
Machine Language

• Why is programming in machine language difficult?
• Must remember the binary encoding for
  instructions or constantly look them up from some
  reference sheet
• Difficult to read code.
• Error prone, difficult to locate errors.
• Recall Brookshear Machine Instructions
• 156C
• 166D
• 5056
• 306E
• C000

8
Machine Language

• Early programmers used machine language to
  program
• plan and lay out data and code in memory
• In BSML, recall starting program at 00 and
  storing data at say 20
• use explicit addresses for jump instructions
• In BSML, recall we had to specify the address to
  jump to
• use explicit addresses for data storage
• What happens if you add another instruction in
  the middle of your code?
• It might move the rest of the instructions in
  memory which
• changes the address of other instructions which
  
• might require changes to the addresses in jump
  instructions.
• Similar problems occur if you change the
  arrangement of data.

9
Assembly Language

• Uses Mnemonic codes
• Assigns Names for memory locations
• Depends on the Computer
• Needs an Assembler to translates from Assembly to
  machine language so the machine can execute the
  program

10
Assembly Language

• Humans work better with mnemonics, which are
  words that are designed to aid in remembering
  information.
• Programmers invented assembly language which uses
  mnemonic symbols rather than binary to represent
  machine instructions and registers

11
Assembly Language

• Mnemonic names
• Example,
• assign the name Price to location 6C,
• assign the name ShippingCharge to 6D
• Assign the name TotalCost to 6E
• Then, the following two columns are equivalent
• 156C Load R5, Price
• 166D Load R6, ShippingCharge
• 5056 Add1 R0, R5, R6
• 306E Store R0, TotalCost
• C000 Halt

12
Assembly Language

• Mnemonics still dont solve the problem of
  relocating addresses when code or data is shifted
• Also, data is still referred to by its address in
  memory, we just use a name for the memory cell
  instead of a number
• Symbolic addresses were developed to refer to
  memory locations
• Example, Use of Labels to mark JUMP statements

13
The Assembler

• Programmers used text files containing assembly
  language to describe programs (known as Source
  code).
• Assembly language used mnemonics and symbolic
  addresses
• Assemblers are software tools that translate
  source code (text) into machine language (binary)

14
Assemblers

• Assemblers translate the mnemonic codes to
  binary codes
• Machine instruction is different for different
  machine architecture
• Recall RISC and CISC Instruction sets?
• So, Assemblers must be written for every
  different type of machine
• Programmers need to think in terms of machine
  instructions
• Programs are not portable

15
High Level Languages 3GL

• Identify a collection of high-level primitives
• e.g. loops and selection constructs
• Think and express concepts at a higher level of
  abstraction than machine language
• Implement translation of such high-level
  primitives into machine language using an
  intermediary (Compilers)

16
3GL Third Generation Languages

• 1960s
• Languages designed for specific types of problems
  and used syntax familiar to the people in that
  field
• FORTRAN math
• COBOL business
• Compiler translates from high-level language to
  machine language

17
3GL Third Generation Languages

• FORTRAN
• 1954
• Represent complex mathematical formulas
• C/C has replaced FORTRAN
• COBOL
• 1959
• Business
• Large complex data files
• Formatted business reports

18
Third Generation Languages FORTRAN
FORTRAN
19
Language Translation Interpreters Compilers

• High level languages were invented to
• break away from machine-dependence
• allow the expression of abstract concepts
• Compilers and interpreters are software tools
  that translate high-level source code (text) into
  machine language (binary)

20
Language Translation

• Interpreters
• read each line of text
• translates the text into machine language
• executes the line before reading the next line
• Compilers
• read all text
• translate everything
• create executable files that represent the source
  code

21
Programming Paradigms

• Procedural (Imperative)
• Cobol, Fortran, C, Pascal
• Focused at expressing an algorithm
• Declarative
• Prolog, SQL
• Focused at expressing the solution to a problem
• Functional
• Scheme, Lisp, ML
• Focused at identifying functional components
• Object-Oriented
• Smalltalk, C, Java
• Programs are sets of cooperating components that
  contain both data and operations

22
Programming Paradigms Procedural

• Procedural the development of a sequence of
  commands that, when followed, manipulate the data
  to produce the desired result
• Our Algorithms discussion is based on this
  paradigm
• Find an algorithm to solve the problem and
  express that algorithm as a sequence of simple
  and unambiguous commands
• C, FORTRAN are examples of procedural paradigm

23
Programming Paradigms Declarative (Logic)

• Declarative design and state the problem
  precisely, rather than design a solution
• The trick is to discover and implement a general
  problem-solving algorithm so that
• all you have to do is state the problem in the
  correct format and it already has a solution
• Useful in simulating systems (economic,
  geographical, political) in order to test
  hypothesis
• The general solution is to recompute the value of
  certain key parameters (like GDP, temperature
  etc) over passage of time
• Then design the problem at hand precisely by
  defining the correct parameters
• Prolog is an example of declarative paradigm

24
Programming Paradigms Functional

• Functional View the process of program
  development as connecting predefined black
  boxes
• each of which accepts input and produces output
• in such a way as to produce the required
  input-to-output relationship.
• Mathematicians refer to such boxes as
  functions and hence the name.
• LISP, ML (MetaLanguage), Hope and Scheme are
  examples of Functional paradigm
• Primitives in this paradigm are elementary
  functions which can be used to create more
  complex functions as needed to solve the problem
• Example Compute the average of a given set of
  numbers

25
Programming Paradigms Functional

• Example Compute the average of a given set of
  numbers
• The three boxes or functions are
• Sum, Count, Divide
• Both Sum and Count functions accept the same
  input whereas Sum box computes the sum, the
  Count box keeps track of the count
• The output from these two boxes are the input
  for the third box Divide, which produces the
  desired output.
• In LISP, the above program is written as
• (Divide (Sum Numbers) (Count Numbers))

26
Evolution of programming paradigms

• AI The functional and logical paradigms were
  very popular a decade or so ago as a result of
  the Japanese 5th Generation initiative,
• which adopted a logic programming language as its
  focus in an attempt to use parallel architectures
  to solve AI problems.
• At that time, groups in the UK and the USA
  emphasized the functional paradigm as an easier
  way to program parallel architectures

27
Evolution of programming paradigms

• OOP The object oriented paradigm is now much
  more fashionable - and is more widely used in
  industry today
• One reason for this is because nearly all object
  oriented languages are quite close to the
  imperative paradigm - though this need not be the
  case - and the change from imperative programming
  to object oriented (hereafter OOP) programming is
  less of a shock for programmers brought up on
  imperative programming languages.

28
OOP Languages

• It may appear that, having learnt C, C is the
  obvious choice of the OO languages.
• The problem with C is that it incorporates some
  OO ideas, but not necessarily in a particularly
  clean way
• As a result it is not an elegant language, and
  nor is it a good example of the different
  paradigm. C also derives a lot from C, but is
  just too new.
• Fortunately Java borrows quite a lot from C as
  well and seems to be easier to learn for
  beginners in OOP.

29
Programming paradigms at a glance
30
Traditional Programming Concepts (TPC)

• Variables and Data Types
• Describe data and data format
• Data Structures
• Complex data types
• Constants and Literals
• Using a value directly
• Assignment Statements
• Assign values to variables
• Assign result of an evaluation to a variable
• Control Statements
• Selection
• Loop

31
TPC Variables and Data Types

• Variable is a location in main memory referenced
  by a name rather than a numeric address
• Variable holds a value value associated with
  that name can change during the program execution
• Variables need to be declared before they can
  be used in the program
• Variable declaration includes specifying what
  type of quantity it can hold numeric,
  character, Boolean and so on known as the data
  type
• Variable declaration in C
• float price
• int age
• char x

32
TPC Data Structures

• Data Structure refers to the conceptual
  arrangement of data for easier manipulation,
  depending on the computation
• Arrays homogenous block structure of cells, with
  each cell in the array holding a value of the
  same data type can be 1D or 2D
• Linked Lists a list of boxes linked by pointers
  to the next box can be doubly linked, circular
  linked
• Trees conceptual arrangement with a single root,
  with parent-child relationship between nodes
• In JS, a one-dimensional array (1D array) is
  written as arrayNameindex. Remember?
• In C, a 2D array holding integers can be declared
  as
• int Scores 2 9
• To be interpreted as a 2 rows and 9 column table
• More on these special data structures when we do
  chapter 7

33
TPC Constants and Literals

• Literal is a fixed quantity that is used over and
  over literally in a computation.
• In calculations of air traffic in the vicinity of
  an airport, if the altitude above sea level is
  645 feet and is a quantity that is used over and
  over in calculations, it is used explicitly as
  is.
• For example,
• EffectiveAltitude AltimeterReading 645
• Where EffectiveAltitude and AltimeterReading are
  variables

34
TPC Constants and Literals

• Constant is a fixed quantity that is referenced
  by a descriptive name rather than the literal
  value
• In calculations of air traffic in the vicinity of
  an airport, if the altitude above sea level is
  645 feet and is a quantity that is used over and
  over in calculations, we can declare it as a
  constant
• Constant AirportAltitude 645
• If EffectiveAltitude and AltimeterReading are the
  same variables as before, then, the same formula
  now can be written as
• Effective Altitude Altimeter Reading
  AirportAltitude

35
TPC Constants and Literals

• Why is constant a better way to program than
  literal?
• Consider the case where a program uses the
  airport altitude 25 times in its calculations.
• Literal method would have the literal value 645
  appearing 25 times in that program
• Constant method would have the value 645
  appearing only once at the initial declaration
• Constant AirportAltitude 645
• Any subsequent use of this value would be done
  using the variable AirportAltitude, not the
  literal value 645.
• So far so good.
• What if we want to use this program with another
  airport with airport altitude 1000 feet?
• Literal method would involve making 25 changes,
  whereas Constant method would involve making only
  one change.

36
TPC Assignment statements

• Assign a right-hand-side (rhs) quantity to a
  left-hand-side (lhs) quantity.
• Z X Y
• rhs quantity X Y is evaluated first and the
  result is assigned to the lhs variable Z
• Name Ann
• The value Ann is assigned to the variable Name
• Many High level languages allow arithmetic
  assignment of the form
• variable operator value

37
TPC Control statements

• Imperative statement that alters the execution
  sequence of a program
• Selection
• If Else, or multi-way switch
• True/false condition to select the set of
  instructions that need to be executed
• Repetition
• While () and for( ) loop structures
• Initialize, modify, terminate the condition to
  control the repetition of a set of instructions
  as needed

38
Procedural Units

• Procedure is an abstract modular tool that can be
  used by several different program units.
• A Procedure needs to be defined first, and then
  invoked via procedure call wherever needed
• Control transfers from current statement in the
  program unit to the procedure unit by means of a
  JUMP control at the time of procedure call
• Control is returned to the original program unit
  once the procedure unit has finished executing
• Procedures have their own set of local variables
• Values from one program unit to another procedure
  unit can be passed via parameters declared for
  that procedure unit
• Reading Chapter 5 Section 5.3

39
Procedural Units Control transfer
40
Procedural Units Procedure Call
41
Procedural Units Separate copies maintained
42
Procedural Units Preserve calling environment
43
Procedures vs. Functions

• Function is similar to Procedure except that a
  value is transferred back to the calling program
  as the value of the function
• Procedures do not pass back a return value and
  preserve the calling environment
• Functions pass back a return value that can be
  used by the calling program whenever needed

44
Grammar

• Every language has a distinct grammar, which is a
  set of rules for creating expressions in the
  language
• (define (sum n m) ( n m)) in Scheme
• x n m in C
• The syntax of a language consists of the grammar
  combined with an alphabet
• Most languages use the normal alphabet, e.g.
  Scheme
• Some languages use special symbols, e.g. APL

45
Syntax Diagram

• Pictorial representation of programs grammatical
  structure
• Example ifthenelse structure
• Terms in rectangles require further description
  and are called non-terminals
• Terms in ovals are called terminals (a.k.a Key
  words or Reserved words)

46
Language Syntax

• Syntax is usually represented in a formal way
  using what is known as Context Free Grammar (CFG)
• One of the widely used formalism is called BNF
  Backus-Naur form named after the originators of
  this formalism

47
Syntax BNF

• The valid and legal syntax of a language can be
  described using a special notation known as
  Backus-Naur form (BNF)
• John Backus and Peter Naur
• Originally developed to describe ALGOL language
  (1950s and 1960s)

48
BNF

• The meta-symbols of BNF are
• meaning "is defined as"
• meaning "or"
• lt gt angle brackets used to surround category
  names.
• The angle brackets distinguish syntax rule names
  (also called non-terminal symbols) from terminal
  symbols which are written exactly as they are to
  be represented.

49
BNF

• Consider a small child that can only form the
  sounds from its limited vocabulary just using M,
  B, D, A, O
• In BNF form, this can be defined as
• ltconsonantgt M B D
• ltvowelgt A O
• ltbaby-talkgt ltconsonantgt ltvowelgt
• In English, the above three BNF statements read
• Consonant is (defined as) M or B or D
• Vowel is (defined as) A or O
• Baby-talk is (defined as) a consonant followed by
  a vowel

50
BNF

• From this specification you can assume the child
  can make the following set of baby-talk
• MA MO BA BO DA DO
• baby-talk ltconsonantgt X ltvowelgt
• Suppose the child is capable of a brief pause
  between sounds PAUSE, we can define
• ltbabblegt ltbaby-talkgt PAUSE ltbabblegt
  ltbaby-talkgt

51
Date Syntax in BNF

• Purpose Write BNF
• Who Groups of 2-3
• Task Write a BNF specification for date syntax
  dd-mmm-yyyy in the form 04-Nov-2002
• Allow illegal but syntactically correct dates
  (e.g. 93-Feb-2999)
• Date is always expressed with two digits and year
  with four digits, while month uses 3-letter
  abbreviation
• Product Alphabet and BNF specification of the
  syntax
• Time 5 minutes

52
Date Syntax BNF

• Possible Solution
• ltmonthgt JanFebMarAprDec
• ltdigitgt 0123456789
• ltddgt ltdigitgtltdigitgt
• ltyyyygt ltdigitgtltdigitgtltdigitgtltdigitgt
• ltdategt ltddgt - ltmonthgt - ltyyyygt

53
OOP vs. Functional JS vs. Scheme?

• Students interested in learning more about
  functional paradigm using Scheme programming
  language can use the information posted on the
  class web page for a self-paced study.
• We will learn a little more about OOP and
  introduce you to JavaScript (JS) to help you do
  your group project.

54
Scheme Syntax

• ltlettergt ltagtltZgt
• ltdigitgt 1230
• ltintegergt ltdigitgtltintegergtltdigitgt
• ltrationalgt ltintegergt / ltintegergt
• ltinexactgt i ltintegergt . ltintegergt
• ltbooleangt true false
• ltvargt ltlettergt ltvargt ltlettergt
• ltcongt ltintegergt ltrationalgt ltinexactgt
  ltbooleangt
• ltprmgt - .

55
Scheme Syntax

• ltexpgt ltvargt
• ltcongt
• (ltprmgt ltexpgt ltexpgt)
• (ltvargt ltexpgt ltexpgt)
• (cond (ltexpgt ltexpgt) (ltexpgt ltexpgt))
• ltdefgt (define (ltvargt ltvargt ltvargt) ltexpgt)
• More on this later

56
Compiler Parse Tree

• The compiler will build a parse tree to represent
  the program based on the grammar
• A parse tree is a graphical representation of the
  syntactical structure of a statement

57
Example Syntax Diagram and Parse Tree for
Expression x y z
58
Compiling a Program

• The steps completed by a compiler are
• Lexical Analysis identify the individual tokens
  in the language statements reading symbol by
  symbol
• Syntax and Semantics Analysis Build a parse
  tree by reading the tokens analyzed by Lexical
  analyzer and Identify the intent of the program
  statements
• Code Generation Produce machine language
  instructions

59
Program Translation Process
60
Program Translation Parsing

• Early parsers required fixed-format for source
  code
• Modern 3GL are free-format languages
• Example
• if cost lt cashOnHand pay cash else use
  credit card
• is understood by the modern parsers correctly,
  but, makes it harder to read for humans, so, we
  write it as
• if (cost lt cashOnHand)
• pay cash
• else
• use credit card
• Semi-colons are typically used to indicate end of
  each statement
• Key words are defined which cannot be used in
  programs to represent anything other than what
  they are defined for

61
Choosing a Language

• Choose based on
• What is available
• Required interface
• What do you know best?
• More importantly Which language lends itself
  best to the problem at hand?

62
Some of the widely used 3GL

• Visual Basic
• 1987
• Create complex user interfaces
• Uses standard Windows features
• Event-driven user controls the program
• C
• 1972
• Efficient code
• Portability
• C
• Enhancement of C with OOP concepts

63
What about Java?

• The Hype about Java
• Java is a simple but powerful language for
  creating safe, portable, robust object-oriented,
  multithreaded programs.
• So, what does it all mean?
• Simple and Powerful
• It should be easy to learn, but you can achieve a
  lot. There should be no surprising features, and
  no fiendishly complicated ones either.
• Safe and Robust
• Java is intimately connected with the Internet.
• you can use Java to write programs which run on
  other peoples computers when they read your Web
  page.
• In such circumstances, programmers and users
  want assurance that bugs in the code cannot
  accidentally do damage to the remote computers on
  which it is run, and that a malicious programmer
  cannot deliberately do damage.
• So a lot of thought has gone into making the
  behavior of programs predictable, and being able
  to run them securely.

64
Java

• Portable
• i.e. it must be architecture neutral.
• It is very easy to write C programs which can
  discover the word-length of your machine, the
  precision of your floating-point representation,
  etc. These are not architecturally neutral!
• Another issue related to portability is character
  sets. Java wants to be able to represent
  information in any language - and therefore does
  not want to be tied to a character set associated
  mainly with English. A character in Java is
  encoded using Unicode, which uses 16 bits per
  character!
• Aside
• Java is Multithreaded
• Not many languages have threads built-in, but
  such concurrency is very useful if you are trying
  to provide a graphical user interface (GUI) to an
  application or if you are communicating with
  another computer. It is possible to add library
  features to C for this sort of thing, but then
  the responsibility for handling the interactions
  between threads is usually just an added burden
  on the programmer. In Java threads have been
  there from the start - and are therefore easier
  to manage.

65
4GLVery High-Level Languages

• Programmer specifies the desired results the
  language develops the solution
• Many times more productive with a 4GL than a
  procedural language
• Query Languages
• Retrieve information from databases
• Easy to learn and use

66
5GL Natural Languages

• Resemble natural or spoken English
• Translates human instructions into code the
  computer can execute

67
OOP Object-Oriented Programming

• Object
• Self-contained unit of data and instructions
• Includes
• Related facts (data)
• Related functions (instructions to act on that
  data)
• Example
• Object cat
• Data feet, nose, fur, tail
• Functions eat, purr, scratch, walk
• Cat Kitty, Susan
• More on OOP in another lecture presentation

68
Summary

• All programs must be translated into machine
  language in order to run
• There are at least three identifiable programming
  paradigms, each with its own special features
• High level languages provide machine independence
  and better productivity for developers and need
  some intermediary like compliers to translate to
  machine language
• Languages are specified using grammars. This
  allows compilers to translate the programs
  unambiguously by creating appropriate parse trees
• Traditional programming concepts include
• Variables, Data Types, Data Structures
• Assignment Statements
• Control Statements like Selection and Repetition