Say you see a puzzle in the newpaper which looks like this:
--- 87- ---
--1 --4 ---
5-- 2-- 3--
2-- -9- -6-
43- --- -25
-9- -3- --7
--8 --5 --9
--- 1-- 2--
--- -69 ---
(Where "-" means a cell with nothing in it)
You can enter the puzzle by putting the cursor on r1c4 and pressing 8. The cursor will automatically go to r1c5 which is logical, right?
OK, so you enter 7 and now you use the mouse again to go to r2c3 in order to enter the 5. Etc, etc.
Well, I am a pretty good speed typist on the numeric keypad. So I would not mess with the mouse but instead would start out
---87------1--4---5--2--3--, etc.
The zero is nice and easy to hit on the numeric keypad, so I use it to skip boxes. Actually I would enter
000870000001004000500200300, etc.
So the zero is intended to enter nothing. The spacebar does that, too and sometimes I use it when I am solving and thus not using speed typing.
For your other question: 213 was just the max I hit. But it would be interesting to compute the worst case. Assume a typical puzzle that has about 25 cells originally filled in. We have 81-25 cells to go. Further assume that rather than entering 1-9 in order, we enter them at random and, unfortunately, the last one entered is always the correct one. (I'm assuming we don't try the same number twice in a given cell)
So we will have 8x56 = 448 errors. That would be the max although my program doesn't provide for a max. I guess if you continually and stubbornly guessed 8 in cell r1c1 long enough, the program will bomb out on the limit of BASIC's INTEGER range. I know I didn't put in a check, so it is one of the few programs I publish that allow the user to crash the program.
Maybe it is the only one.
Mac