Arotau

Today's work

• Dug up the old NR source code.
• Put my thesis into revision control - lets me automatically check on progress each day.
• Got all the citations in the interaction chapter sorted out.
• Went through the results of the most recent assignment table experiments, and found that the preprocessor was proving a number of satisfiable instances unsat. Eventually tracked it down to the equivalence checking code. My first guess was that the equivalence detection was OK, and that the replacement of one variable with another was flawed. However, it appears that the logic for finding the equivalences is incorrect, though I still haven’t figured out exactly why.