DSM uses metrics that are based on the speci?¬?c domain and thus uses the metamodel
data. For example, a metric for software used to control production processes in a
paper mill can calculate valves, motors, pumps, and their characteristics rather than
produce general metamodel-independent metrics on system complexity (e.g.,
cyclomatic complexity by McCabe, ) or program length (Halstead, 1977).
Prototyping and Simulation Generators can also be used to produce
prototypes instead of production code. Prototyping is usually used to give early
feedback, and generators may produce code for a totally different target platform and
in a different programming language than the production code. Here generators don??™t
necessarily need to optimize the code, but to enable functionality, usability, look and
feel or other characteristics relevant to prototyping. The modeling language, however,
can be the same for developing both the prototype and the production code. The model
can also be used for simulation: code generators then provide output in a syntax
required by a simulator.
Con?¬?guration, Packaging, and Deployment Generators can be limited to
producing component glue code, such as integrating existing components together.
When seeking better automation, a more typical case is to produce con?¬?guration
data and other packaging information for application deployment and installation.
Pages:
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183