If you wish to skip the preamble go here.
A pixel is the smallest visible unit a digital photograph can be split up into. In a photograph they are arranged in regular rows and columns. You are probably familiar with X and Y coordinates, for instance on a graph. Individual pixels can be identified by their coordinates. Because CombineZ is a Windows based program, it follows the Windows coordinate system: the upper left Pixel has coordinates (0,0), X increases as you move to the right, and Y as you move down. This notion can be extended one step further; each Frame in the Stack can be given a number that becomes its Z coordinate, which increases as you go up the stack. Thus (1,2,3) refers to the second pixel along the third row down on the 4th Frame. Remember these values start at zero.
Each Pixel in the finished Picture only has one colour, but at its (X,Y) coordinates there are as many Pixels to choose from as there are Frames in the Stack. If you were to take each Pixel from the Frame that appears to be most sharply in focus at each coordinate, your Picture would look pretty good. Remember each frame is assigned a number, Z. If you were to collect the Z numbers of every Pixel in a Picture that was formed using the above rule, and store them in an array in the right positions, then they too would have X,Y coordinates that correspond to the location of their pixel in the finished Picture. This array is called a Depthmap.
You can reconstruct a Picture from the Stack and a Depthmap by visiting each (X,Y) pixel position in turn, looking up its Z value in the Depthmap, then getting pixel (X,Y,Z) from the Stack and placing it on the Picture. CombineZ goes one step further; it allows Z to have values which lie between two integers. For example Z=1.5 means the Pixel should come from a point which lies half way between Frames 1 and 2, thus its value is calculated to be halfway between those two Frames' values.
Find Detail, and Patchwork, look for the most in-focus pixels at each (X,Y) location. Frequently they can't decide so they leave a Pixel in the Picture empty. Sometimes they make a mistake and find the wrong Z value, but (hopefully!) most of the time they get it right.
Fill Gaps takes any pixels that are there and uses them to estimate what the missing pixels should look like by predicting their Z values.
Both of these steps are prone to make errors. Noise in a Frame can be misinterpreted as Detail. Missing values are only filled in with a calculated guess; if noise is present this can be wildly wrong. So you see it is quite often necessary to correct mistakes made by the program. You can do this externally in a program with which you are familiar, or internally in CombineZ.
If there is a large expanse without detail, or one which appears to have detail in a frame due to the spread of an area of a different colour as it goes out of focus, you can try 'Seeding' this area, i.e. put some Pixels from the right Frame there and Fix them before running Find Detail. Then when you Fill Gaps there will be some more correct Pixels to base the guesswork on.
You may find it necessary at times to include Pixels from the wrong Frame for appearances' sake; in fact it is often essential where contrasting coloured areas meet that come from different Frames. The reason for this is not difficult to understand. When the background is in sharp focus, the foreground Frame is out of focus and the area in question from the foreground will show on the background Frame because things spread as they go out of focus.