Finally, after the above-mentioned theoretical studies and experimental work have been completed, I will spend the Fall semester writing up the results in the thesis document. In addition to relating the work done as described above, the thesis document will, as time permits, discuss a variety of related issues, many of which were mentioned in the earlier, longer version of this proposal (available at http://www.ai.mit.edu/mpf/proposal/proposal.html).
For example, there are possible applications of reversibility for error detection, debugging, and auditability which should be discussed. There are also questions about how to manage multiple, inter-communicating reversible processes within a multiprocessing framework, and about how to relate reversibility to rollback in transactional database systems.
Also, there is the very interesting issue of how to extend computational complexity theory to incorporate facts of physical reality other than just the second law of thermodynamics, such as, notably, the finiteness of the speed of light (which traditional complexity theory conveniently ignores by assuming that any of an unlimited number of RAM cells can be accessed in constant time). The longer version of the proposal mentions numerous other interesting issues as well.
However, all these tangential discussions will be optional, and will be undertaken only after the main core of the thesis research, proposed here, has been successfully completed and documented. If no time remains to pursue these other directions, the coverage of these topics in the thesis will be limited to the short discussions already written up in the longer version of this proposal.