Short note: generator-based reuse

Generator-based reuse is an approach in which reusable knowledge is captured in a program generator instead of in code components. The developer describes what is needed in a domain-oriented language or through an interactive tool, and the generator automatically produces the program code.

+-------------+    +-----------+    +-----------+
| Application |--->| Program   |--->| Generated |
| description |    | generator |    | program   |
+-------------+    +-----------+    +-----------+
                         ^                ^
                         |                |
                 +-------------+  +--------------+
                 | Application |  | Extra hand-  |
                 | domain      |  | written code |
                 | knowledge   |  | (if needed)  |
                 +-------------+  +--------------+

Examples

• Application generators for business data processing (report and form generators).
• Parser and lexical analyser generators such as yacc and lex, which generate a parser from a grammar.
• Code generators in CASE tools that produce code from UML/design models.
• GUI builders that generate interface code from drag-and-drop screens.

Advantages

• Very cost-effective: a short description produces a lot of correct code.
• Easier for domain experts and end-users than writing code.
• Generated code is consistent and follows standards.

Disadvantages

• Applicable only to well-understood, limited domains where abstractions can be clearly mapped to code.
• Generated code may be less efficient and hard to read or modify by hand.
• Building the generator itself is expensive.