The need for, and the characteristics of, translators (assemblers, compilers, interpreters), the stages of compilation (lexical analysis, syntax analysis, code generation, optimisation), and the role of linkers, loaders and libraries.

Award up to two marks each for error reporting, speed and distribution.

Error reporting: a compiler translates the whole program at once and reports all syntax errors together after the attempt, so debugging happens in batches; an interpreter translates and executes one statement at a time and stops at the first error, which is convenient for development and locating errors.

Execution speed: compiled code runs as native machine code, so it executes quickly with no translation overhead at run time; interpreted code is translated each time it runs, so it is slower, and a statement inside a loop is retranslated on every pass.

Distribution: a compiler produces a standalone executable that can be distributed without the source code (protecting it) but is platform-specific; an interpreter needs the source code and the interpreter present on the target machine, which aids portability but exposes the source. Markers reward a clear point on each of the three aspects.

Award marks for each stage described in order (up to 5).

Lexical analysis: the source code is read and broken into tokens; whitespace and comments are removed and a symbol table is begun to record identifiers and keywords.

Syntax analysis (parsing): the tokens are checked against the rules of the language grammar, building a structure such as an abstract syntax tree; syntax errors are reported here and the symbol table is completed.

Code generation: the syntax tree is translated into object code (machine code or an intermediate form). An optional fourth stage, optimisation, then improves the code to run faster or use less memory, for example removing redundant instructions. Markers reward tokens at lexical, grammar checking at syntax, and machine/object code at code generation.