Toolbar and menus
A toolbar and two menus provide access to commands and configuration options. An information bar gives data on the state of play for the current puzzle.
The Toolbar is a set of buttons which execute the most common program commands: Undo(<), Undo until consistent(<<), Give hint(?), Apply simple filter exhaustively(s), Simple filter(S), Impossible combination(IC), Unique intersections(UI), Unique candidate's combination impossible(UC), Hidden singles(H1), Hidden pairs(H2), Hidden triples(H3), Naked pair(N2), Naked triple(N3), Naked quad(N4), Xwing(XW), Possible permutations(PP), Candidate bounds(CB), Isolated necks(IN), Best guess(BG).
Toolbar options are activated by clicking with the left mouse button. If a hint is requested and found the "?" button and the button for the found hint will be shaded red. The hint can be executed - i.e. the red candidates deleted - by left-clicking on either of the red buttons in the Toolbar. The hint can be cancelled by right-clicking on either red button in the Toolbar. Clicking on a Toolbar button with the middle mouse button will bring up a pop-up which describes the function of the button. An example is shown in Figure 2.
The File menu provides access to new puzzles, checks puzzles for having a unique solution, controls the difficulty rating system, controls the help and hint status, sets cell and candidate colours, sets the symbols used and the current hint search strategy.
The Edit menu contains exactly the same options as those available from the Toolbar.
The Information bar
The five numbers at the base of the puzzle grid give the Puzzle Number, the number of steps taken by the user (Clicks), the number of states recorded (History), the remaining "Information content" of the puzzle and the number of seconds (Time) since the puzzle was started.
When the user requests a hint the program searches using one of two strategies. The first, Most effective, finds the hint which can remove the most candidates. The second, Simplest hint, finds the least difficult hint - i.e. the one with the lowest algorithm score. The choice of which one to employ is set using the "Set search strategy" option in the File menu.
The program includes several ways in which the user can recover from possible errors. These include the ability to turn on and off individual candidates and two undo methods.
The program has a simple history mechanism which records every state of the grid. Each time the user unsets a candidate or sets the solution for a cell the history clicks forward by 1. Each time the user employs an algorithm which removes candidates the history clicks on by 1 (irrespective of how many candidates it removes).
The history mechanism allows steps to be undone. This is a one-way process: once a step backwards is taken the deleted operation can only be recovered by being repeated - it is lost to the history mechanism.
When undo is used all candidate and cell colouring, and all hint colouring is switched off.
The candidates for a cell are switched off by clicking on their buttons with the left mouse button. To reactivate a button, left click on the "R" button in the cell - this will switch on all candidates in the cell. When the cell solution is set the "R" will disappear, but the button beneath it remains active and so can be used to reinstate all the cell's candidates.
A single step can be undone by use of the "Undo" option in the Edit menu or the "<" button in the Toolbar. Repeated use will step back through all the states the grid has taken. However, this is one-way only.
As described the program can perform two levels of check on the state of the grid: Is the grid consistent? Have incorrect answers been set? The grid status can be constantly monitored and errors flagged. Confident solvers who work without these automatic checks being applied may reach a state where the grid is in a real mess. In this situation the "Undo until consistent" command available from the Edit menu or "<<" in the Toolbar is quite useful: it will step back through the grid history to the last state in which the grid was consistent and contains no incorrect answers.