Language Evaluation

Introduction - Why 380??

1. Increased capacity to express ideas.
2. Improved background for choosing languages.
3. Increased ability to learn languages.
4. Understanding of significance of implementation.
5. Ability to design new languages.
6. Overall advancement of computing.

Language Evaluation Criteria

1. Criteria
   1. Readability
   2. Writability
   3. Reliability
   4. Cost
2. Affected by
   1. Simplicity
   2. Orthogonality
   3. Data types and structures
   4. Syntax
   5. Support for Abstraction
   6. Type checking
   7. Exception handling
   8. Aliasing
   9. Control structures
   10. Expressivity

Criteria: Readability & Writability

1. Simplicity
   1. Small number of basic components
   2. One syntax :: one meaning
      • Increment x with x++; x=x+1; x+=1; ++x;
      • Y=SUM(I,J); Array or function ???
      • Perl deliberately violates
2. Orthogonality
   1. Any composition of basic primitives is allowed.
   2. Need small set of primitives and ways to combine them
   3. Violation: Out parameters for functions
   4. Violation: Return types for functions
   5. Violation: Functions as first-class objects
   6. Non-orthogonal often to simplify implementation

Criteria: Readability & Writability (Cont.)

1. Control Statements/Constructs

   Which control mechanisms to include?

2. Data Types

   * Primitive (e.g. boolean, char, unsigned)

   * Structured

   * Provide abstraction.

3. Syntax

   * Matters more than you think!

   * Identifier length, reserved words, layout, etc.

4. Abstraction

   * Hide details to manage complexity!

   * Process abstraction

   * Data abstraction

Criteria: Reliability

1. Performs to specifications under all conditions
2. Impacted by:
   1. Type checking
      1. compile-time
      2. runtime
      3. none
   2. Exception handling
   3. Aliasing (i.e. Two or more names for same memory cell), for example:

      p,q: access Integer;
      
      p := new Integer;
      q := p;
      

Criteria: Cost

1. Programer training
2. Program writing
3. Program compilation
4. Program execution (speed and size)
5. Liability
6. Tool availability
7. Application
8. Programmer availability
9. Support for language
10. Compiler and tool cost
11. Maintenance time (>50%!)

Programming domains

• Science
• Business
• AI
• Systems
• Real Time/Embedded systems
• Scripting
• Web
• Database

Influences on language design

• von Neumann architecture:
  • Memory, I/O, CPU (ALU and CU)
  • von Neumann bottleneck
  • Reflected in languages!
  • Other architectures??
• Programming methodologies
  • Structured programming
  • Procedure abstraction
  • Data abstraction
  • Objects

von Neumann Architecture

1. Fetch-Execute cycle
   1. control fetches next instruction from memory
   2. control decodes instruction
   3. control causes instruction to be executed
      1. move data from memory to ALU, or
      2. move data from ALU to memory , or
      3. do I/O
2. Fetch-Decode-Execute!!
3. Reflected in imperative languages:
   1. Sequential execution of instructions
   2. Modifiable memory locations (variables)

Language Paradigms

• Imperative: Specify how, not what
  • Procedure oriented
  • Object oriented
• Declarative - Specify what, not how
  • Applicative (i.e. functional )
    • LISP
    • ML, Haskell
    • APL, FP
    • Quicksort in Haskell
  • Logical (i.e. rule based) - Prolog

Evolution of Imperative Languages

• Control abstraction
  • Statement level
  • Structured statements
  • Procedural level
  • Threads
  • Exceptions
• Data abstraction
  • Records
  • Modules
  • Generics
  • ADTs
  • Classes and objects
  • Garbage collection