Typos can make this unreadable, I apologize. Here it is corrected:
Looking through the code for my sudoku solver, I think I found the (or a) 
reason why it wasn't working. I keep track of all unsolved cells' possble 
values. If a cell gets solved, the cells of its face, row, column get updated 
by removing the value of just solved cell  using "e.". Each updated cell's 
face, row, column would then as well need to be updated. I'm currently doing 
this with one global variable holding for the the just solved cell's value, and 
loops, wherein the recursion is: if you remove the, say, "3" from the 
possibilities of "2 3" of one cell in the face or row or column of just solved 
one (whose value is now 3), this global value would assume the value of "2". I 
would like this variable to return to its value of "3" after the call to the 
verb with its value "2" end.

This is the difficulty of working with at most 2 arguments (of the same type), 
right? What does one usually do when encountering such a problem? Pass a box 
with >1 elements and interally assign private variables eachtime, and on each 
necessary recurssion pass everything as ordered list of boxes, or is there a 
way of editting the scope of a to-be-called verb (or part of sentence)?
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm

Reply via email to