It's more complicated than just looking at dots/no dots. You also have to take into account the pattern of continuous diagonal dots. Having no dots can mean a substitution. Usually a shift of the diagonal indicates a deletion or insertion.

First, you need to tell us if this is seq A dot-plotted against seq A' or against seq B, where A' is the version very much like A but with suspected insertions/deletions, but B is simply a much more distantly related sequence.

Second, one needs to know the window and threshold for plotting the dot. If the window = 1 and the threshold = 1, then perhaps a dot = a match and no dot = in/del - or no dot = a mismatch. Normally the window is much greater and the threshold is less than 100% (match).

Thus, as DK writes (+1), it is more complicated than dot/no dots. As I often tell people, many such bioinformatics activities are easy. So, just set up a series of A type sequences, where you have seq A, A + insertion, A - deletion, A + A, rev complement of A and so on. Then, make the dotplots of all these against A to see what they look like. Once you've done that, you'll have your own dotplot glossary, but more importantly, you'll have taught yourself much more than we can type here and you'll be smarter for it.