InternalRect: [L2, T2, R1, B1], ExternalRect: [L1, T1, R2, B2].
If the user sets no constraints for the boundaries of the search area, the internal rectangle degenerates and makes no geometrical sense. In this case L1=R1, L2=R2, T1=B1, T2=B2.
Suppose now that no constraints have been specified for the search area. Then, if the maximum and minimum allowed number of pages is 1, the fuzzy rectangle of the search area will be created as follows:
Operations on Fuzzy Rectangles
Fuzzy rectangles can be united and intersected. When two fuzzy rectangles are intersected, their corresponding boundaries intersect: the right boundary of one rectangle intersects with the right boundary of the other, the left with the left, etc. This creates new ranges for the boundaries:(L1 <= L2, T1 <= T2, R1 <= R2, B1 <= B2) is not met, the resulting fuzzy rectangle will be nullified: InternalRect: [0, 0, 0, 0], ExternalRect: [0, 0, 0, 0]. Usually an empty fuzzy rectangle is created as a result of contradictory conditions, e.g. if you attempt to intersect two fuzzy rectangles whose external boundaries do not intersect.
Fuzzy rectangles are united in the same way: first, the ranges of each boundary are united. As a result of uniting the two ranges, a minimum range is created which contains both specified ranges. After the boundaries have been united, a new fuzzy rectangle is formed.
In practice you will most often encounter fuzzy rectangles with a degenerated internal rectangle. Only on rare occasions will you need to set constraints on a specific boundary. However, the search area for an element is always represented by a fuzzy rectangle and an array of excluded rectangles.