Designing a reusable symbol table library

Presents the design of a symbol table library and addresses its design rationale. A symbol-processing task can be divided into two parts: one for storing/retrieving symbols, and the other for semantic checking. The former is language-independent and can be defined in terms of reusable components. The latter is language-dependent and can be defined in terms of parametric types of symbol table components or as specific semantic rules. We identify general properties of name analysis problems: topology, overload, navigation, etc. The resulting design of the library covers most usages in symbol processing.<
>