159'331 Programming Languages

1
159.331 Programming Languages Algorithms

• Lecture 03 - Aspects of Programming Languages -
  Part 2

2
Programming Paradigms

• Paradigm - unfortunately word has become a bit
  abused
• From Greek
• para - side by side and
• deigma - that which is shown
• In (Computer) Science a paradigm is usually
• A coherent set of methods that have been found to
  be (more or less) effective in handling a
  particular type of problem (problem domain)

3
Defining Paradigms

• Usually a paradigm can be described by a simple
  core guiding principle
• Beware however that not all groupings or models
  that are declared to be paradigms are equally
  coherent
• Sometimes paradigms in science only become
  apparent post hoc (afterwards when the fuss has
  died down)
• We will follow Bal Grunes Programming
  paradigms structure

4
Prominent Programming Paradigms

• Imperative Paradigm
• Object-oriented Paradigm
• Functional Paradigm
• Logic Paradigm
• Parallel and Distributed Paradigm
• Be aware others do exists and are (constantly)
  being proposed and argued for/against
• See the programming literature

5
Programming Models

• Paradigm will be our major sort key
• We will follow Bal Grune and use Model as
  our next subdivision
• A Programming Model therefore is a way of
  dividing up a paradigm for sub problems or areas
• For example we will discuss the message-passing
  and data-parallel programming models that could
  be said to be approaches to the parallel and
  distributed programming paradigm.
• Remember this is only a convention for
  discussing the field - useful if we try to be
  consistent!

6
Bals Grunes way of thinking about a
programming paradigm
7
Some More Standard Jargon

• Generations in computer languages
• Nice idea but lots of controversy about exactly
  what belongs in which category
• Roughly sequential idea
• 1st generation is lowest level, earliest
  generation
• Higher generations or layers are more
  advanced/sophisticated

8
procedural sometimes called imperative
9
Generational Language Development

• We have discussed the first three generations -
  which seem mostly agreed upon
• Fourth Generation Languages (4GL) have
• Strong data structure support with persistence
• Support for application/problem-specific ideas
• SQL is a good example - Structured Query
  language - a programming language for managing
  databases ina platform/vendor independent way

10
5th and 6th Generations

• 5GL are intended to allow non-specialists to use
  Artificial Intelligence (AI) ideas - many based
  on logical Inference
• Prolog is a classic example - it provides a
  logic inference engine with a command line
  interface and language syntax
• 6GLs are a bit less well defined - typically
  based on neural network engines with an
  appropriate syntax - mostly still experimental.

11
Other Language Classifications

• Can emphasise what is to be achieved rather than
  how it is to be done.
• Languages that emphasise the what are termed
  definitional or declarative
• The functional and logic languages tend towards
  declarative
• Languages emphasising the how called
  operational or non-declarative
• The imperative (or procedural) and OO languages
  tend towards operational

12
Language Levels

• Low, High and Very High Level (see table)
• Not terribly meaningful but good at
  distinguishing between assembly etc at the
  lowest level, through Fortran, C, Pascal and
  friends at High and those using an abstract
  machine layer at the Very High Level.

13
SETL is a Set oriented language with good logic
support
14
Program Structure

• Reminder of some terminology.
• In a natural language we usually think in terms
  of lexical, syntactic and semantic levels.
• Lexical level - concerns the vocabulary (set) of
  words
• Syntax says how the words are allowed to be
  strung together to make sentences
• Semantics level addresses what it means
• This does not quite work for programming
  languages

15
Lexical Structure

• Usually this level is pretty obvious to the
  human reader
• Nouns and verbs etc in natural languages
• Special reserved keywords like BEGIN, END, FOR
  in programming languages even is a word of
  sorts
• Program identifiers - typically allowed to
  start with a letter, and be of some max length
  containing any mix of letters and numbers
  and upper lower case.

16
Syntactic Level

• The rules specify what words are allowed to
  appear where
• eg reserved keywords symbols and identifiers in
  a typical programming language
• Often these syntax rules are designed to be
  close to natural language constructs
• eg ifthenelse construct

17
Contextual Layer

• Programming languages generally have an extra
  layer we can call the contextual layer between
  syntax and semantics
• Think of it as a way of breaking down the
  concept of semantics into a form more easily
  explained in a program
• Need this as bits of syntax mean different things
  in different contexts
• This level is concerned with the feasibility of
  instructions and indicates what actions can be
  performed on a given object

18
Semantics Layer

• The meaning
• In programming languages the semantic level
  describes what the permissable actions mean
• Logic and Functional (declarative tending)
  languages have less distinction between
  contextual and semantics layers.
• These languages are less instruction-oriented
  (operational tending).

19
(No Transcript)
20
Lexical Structure

• Lexical symbols, lexical units, tokens
• Identifiers - (our) chosen variable names eg
  myName
• Keywords - the languages special words - eg for
• Operators - the languages special symbols eg
• Separators - more special symbols eg
• Literals - hard wired constants eg 123, ABC
• Comments - ignored by the compiler
• Layout and White Space - allows break up into
  tokens

21
Syntactic Structure

• A context-free grammar - Extended Backus-Naur
  Form (EBNF)
• Program parsing
• Can use a grammar to produce a sequence of
  lexical symbols which is a program (ie we can
  write programs)
• We can parse a program into meaningful units
  according to a grammar

22
means is defined as means or lt gt
denotes notion names
Note - this is not a very exciting language!
superscript is used to indicate repetition 1
or more times superscript used for zero or
more times
23
The EBNF grammar tells out what we are allowed
to write In this program language.
The grammar also specifies what a compiler or
interpreter can do to make sense of the program
text.
24
We break up the program piece by piece - rule by
rule
Computer programs are in fact very good at
following rules!
25
Context

• Context conditions extra conditions in the
  language definition concerning entities
• Our language might forbid some things
• in an expression both values must be integers
  so 1 a is illegal
• if a value is an identifier it must have a
  value so program n n 1 end is illegal
• We would like the compiler to check these
  conditions (not always done or not always
  possible)

26
Semantics

• What do language notions mean ?
• For example - what will happen when the value of
  an expression is calculated?
• This is the semantics of the expression
• Semantics of each notion combine to define the
  semantics of the whole language
• Semantics can be described formally - but much
  harder than context-free grammars used to
  describe syntax - well beyond our scope for this
  paper.

27
Aspects Part 2 - Summary

• Programming Paradigms
• Programming Models
• Programming Generations
• Programming Levels
• Lexical Layer
• Syntax Layer - context-free grammars and EBNF
• Contextual Layer
• Semantics Layer
• See the EBNF links on the web site.
• Chaper 1 of Bal Grune, Chapter 3-4 of Sebesta
• Next - Programming Language Media and Program
  Construction