Here DSM changes the rules: an experienced developer, usually
within a company, de?¬?nes the automation process and output for the rest of the
developers in that team.
Is the Generated Code Ef?¬?cient? The major argument against generators is
the claim that the generated code cannot meet the strict requirements of size and
runtime ef?¬?ciency that are fundamental issues when developing software for devices
with limited memory and processing resources. When comparing generated code and
manually written code, we should not forget that the compiler performs further
optimization. In one application area, a company conducted an analysis of hand
optimization in speed and memory use in the assembly language produced by a
compiler. After careful analysis of the code, the comparison team did not ?¬?nd any
substantial differences. This study was conducted in an embedded area where code
size and memory use matter. While you may have bad experiences with the quality of
code produced by conventional code generators, they are not valid for DSM. You now
have control and can update the generator if needed. Several examples, presented in
Part III, especially target embedded software development with relatively strict
requirements for the code.
When code is produced by a generator made by an experienced developer, it will
always produce better code than the average programmer writes manually.
Pages:
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180