Exporting a metamodel to a different tool may be useful when con?¬?guring a tool
that is later to process exported models. Currently, exporting a metamodel from one
DSM tool to another has predictable problems with loss of functionality: only the
lowest common denominator is supported by systems such as KM3. For transfer
between current ???version 1.0??? tools this is less of a problem: they have little beyond
this, and what they have tends to be expressed as code anyway, which could not
realistically be transformed. In a situation with several more mature, data-based tools,
better results could be expected from bespoke translations between a pair of tools.
Where a tool supports metamodel import from a textual format, interesting
possibilities exist for turning models into metamodels. A modeling language and
generators can be built that produce this format, and the results can be imported back
into the same tool (or another instance). Many ???version 1.0??? tools indeed use this as
their standard way of creating metamodels, although there is normally some special
handwritten support in the tool for the modeling language in question. In tools with the
power to achieve this without special tricks or hand coding, it provides a rich and
fertile ground for investigating different ways of metamodeling: DSM applied back
on itself.
Pages:
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751