A. Reading the Grid Nomenclature thread I started thinking XML may be
handy for describing Scott's nomenclature. I tried a few examples and
uploaded them to the group. My methods evolved from one example to the
next, so they aren't consistent, but they may serve useful as a
talking point.
B. I don't know what got into me but I was reading this forum the
C. I was just thinking this morning that we may be able to extend
Scott's nomenclature and process to constraint puzzle by allowing a
stream to be the set of all possible solutions for a given fact, then
apply a filter transform that has as its input the list of
possibilities and a constraint generated by the rules of the puzzle.
The output would be another list of possible solutions, probably
smaller. When the list of possible solutions is down to 1, the fact is
fully solved. As an example, one of the streams for a crossword puzzle
may start off as a list of all 5 letter English words, generated by
the 5 empty boxes on the grid. Later in the solution process we may be
able to apply a filter by constraining the third letter to be N.
Eventually we may apply a definition transform with "A light, open,
slender boat that has pointed ends and is propelled by paddles." as an
input and the list can be narrowed down to the single word "canoe."
Later transforms would still need to be applied, say constraining the
second letter to "A." If a filter ever generated a null set, it would
indicate an error somewhere in the process.