. Models are the input for multiple different artifacts. Models not only are the
input for generating code but also can be used for many other purposes, such as
producing documentation, test cases, and con?¬?guration data. What later stages
are used depends on the objectives. Some of the generation targets are
discussed in Section 4.4.
. Models can be expressed in different representational styles. Depending on
the model user, visualization needs, and analysis needs, models can be
expressed in different representational styles, such as graphical diagrams,
matrixes, tables, forms, or text.
. Modeling with DSM is agile. Compared to heavy up-front speci?¬?cation work
before the actual implementation starts, DSM is agile: Only those aspects are
modeled that are relevant. With code generation, we can quickly get response
78 ARCHITECTURE OF DSM
and feedback to the models. If desired, the modeling language can be designed
particularly to add agility to the development. For instance, a modeling
language can ?¬?rst be used partially, to produce a prototype to review the
functionality of the product. Later, the same models can be extended to ?¬?nalize
the speci?¬?cation and generate production code, possibly into a different
programming language from that used for the prototype.
4.3.3 Users of Models
Often we focus on making models for code generation, so the typical language users
are application developers.
Pages:
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170