here - Stefan-Marr.de

9. Conclusion and Future Work
all reflective operations are able to follow strictly the existing semantics of the
memory model, it might integrate well with the OMOP.
A more complex challenge is the integration with the security mechanisms
of the platforms (cf. Sec. 3.3.5). These and similarly involved features will
require careful engineering to preserve the existing semantics and integrate
them with the flexibility of an OMOP.
9.5.6. Formalization
A different research direction is opened up by the unifying characteristics of
the OMOP. Using it as a target for a wide range of concurrent programming
concepts enables their representation in one common framework.
Such a research effort could yield valuable understanding of the concrete
design space of these concepts. On the one hand, this would open up opportunities
to engineer libraries of these concepts that can then be used to
build domain-specific abstractions. On the other hand, it would help understand
the commonalities and variabilities of the concepts, yielding a common
vocabulary. Currently, a major issue in the field of concurrent and parallel
research is the lack of a sufficiently clear distinction between concepts and a
missing common vocabulary, making it hard to relate research from different
eras and research domains to each other. Building these concurrent programming
concepts on top of the OMOP would yield a practically usable collection
of artifacts and might result in the discovery of relevant points in the design
space of concurrent programming concepts that have not yet been explored.
Furthermore, describing all concepts in a common framework could help in
identifying fully declarative representations. Currently, the OMOP is a classic
MOP and the customization of intercession handlers is prescriptive, while a
declarative approach could yield a more descriptive representation. Descriptive
representations could be valuable to synthesize interaction semantics,
which is one of the current limitations. Furthermore, they could become the
foundation for compiler optimizations to improve performance.
9.5.7. Additional Bytecode Set for Enforced Execution
Currently the inherent overhead of the RoarVM+OMOP implementation is
significant enough to prompt the desire for optimization. One possible optimization
would be a separate bytecode set for use during enforced execution.
Thus, instead of checking for the enforcement mode during execution, the
246