Envisaging the
development process of a single product with DSM helps us identify gaps in our
thinking: things that need doing, but for which we have not yet assigned resources.
It may also suggest ideas for further automation of development: phases that are
currently carried out manually by rote, but which the greater precision of DSM would
allow us to automate. Sometimes such improvements may be possible with the
DEFINING THE DSM SOLUTION 343
information captured by the existing modeling language, but more often we will be
able to accommodate them with minor additional properties in graphs.
Perhaps more important is looking at the issues that appear when we have multiple
models. At its heart, this is a question of reuse. By its very nature, DSM has already
provided a massive improvement in reuse and, in fact, has automated this so modelers
do not even need to remember to reuse something. Putting an object in a diagram is
reuse: all the work of the metamodeler in the modeling language, generator, and
domain framework is being reused. Once we have multiple models, however, we open
up the possibility for reuse of models or model elements.
Reuse of models or model elements is similar to reuse in normal codebased
development. Developers need to be aware of what exists that can be reused,
know how to reuse it, and how to make their own components available for
reuse.
Pages:
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647