下一节将通过三个示例说明上述原理。在两个示例中，最好的方法是分离相关的代码段。在第三个示例中，最好将它们结合在一起。
第一个示例由插入光标和 第六章：通用模块更深入 的 GUI 编辑器项目中的选择组成。编辑器显示闪烁的垂直线，指示用户键入的文本将出现在文档中的何处。它还显示了一个突出显示的字符范围，称为选择，用于复制或删除文本。插入光标始终可见，但是有时可能没有选择文本。如果存在选择，则插入光标始终位于其一端。
选择和插入光标在某些方面相关。例如，光标始终位于所选内容的一端，并且倾向于将光标和所选内容一起操作：单击并拖动鼠标将它们都设置，然后插入文本会首先删除所选的文本（如果有），然后在光标位置插入新文本。因此，使用单个对象管理选择和光标似乎合乎逻辑，并且一个项目团队采用了这种方法。该对象在文件中存储了两个位置，以及布尔值，它们指示光标的哪一端以及选择是否存在。
但是，合并的对象很尴尬。它对高级代码没有任何好处，因为高级代码仍然需要将选择和游标视为不同的实体，并且对它们进行单独操作（在插入文本期间，它首先在组合对象上调用一个方法来删除选定的文本；然后调用另一个方法来检索光标位置，以插入新文本）。实际上，组合对象比单独的对象实现起来要复杂得多。它避免了将光标位置存储为单独的实体，而是不得不存储一个布尔值，该布尔值指示选择的哪一端是光标。为了检索光标位置，组合对象必须首先测试布尔值，然后选择选择的适当结尾。
当通用机制还包含专门用于该机制的特定用途的代码时，就会出现此红色标志。这使该机制更加复杂，并在该机制与特定用例之间造成了信息泄漏：对用例的未来修改也可能需要对基础机制进行更改。
在这种情况下，选择和光标之间的关联度不足以将它们组合在一起。当修改代码以分隔选择和光标时，用法和实现都变得更加简单。与必须从中提取选择和光标信息的组合对象相比，单独的对象提供了更简单的接口。游标的实现也变得更加简单，因为游标的位置是直接表示的，而不是通过选择和布尔值间接表示的。实际上，在修订版中，没有特殊的类用于选择或游标。相反，引入了一个新的 Position 类来表示文件中的位置（行号和行内的字符）。选择用两个位置表示，光标用一个位置表示。职位还在项目中找到了其他用途。
The next sections work through three examples that illustrate the principles discussed above. In two of the examples the best approach is to separate the relevant pieces of code; in the third example it is better to join them together.
The first example consists of the insertion cursor and the selection in the GUI editor project from Chapter 6. The editor displayed a blinking vertical line indicating where text typed by the user would appear in the document. It also displayed a highlighted range of characters called the selection, which was used for copying or deleting text. The insertion cursor was always visible, but there could be times when no text was selected. If the selection existed, the insertion cursor was always positioned at one end of it.
The selection and insertion cursor are related in some ways. For example, the cursor is always positioned at one end of the selection, and the cursor and selection tend to be manipulated together: clicking and dragging the mouse sets both of them, and text insertion first deletes the selected text, if there is any, and then inserts new text at the cursor position. Thus, it might seem logical to use a single object to manage both the selection and the cursor, and one project team took this approach. The object stored two positions in the file, along with booleans indicating which end was the cursor and whether the selection existed.
However, the combined object was awkward. It provided no benefit for higher-level code, since the higher-level code still needed to be aware of the selection and cursor as distinct entities, and it manipulated them separately (during text insertion, it first invoked a method on the combined object to delete the selected text; then it invoked another method to retrieve the cursor position in order to insert new text). The combined object was actually more complex to implement than separate objects. It avoided storing the cursor position as a separate entity, but instead had to store a boolean indicating which end of the selection was the cursor. In order to retrieve the cursor position, the combined object had to first test the boolean and then choose the appropriate end of the selection.
This red flag occurs when a general-purpose mechanism also contains code specialized for a particular use of that mechanism. This makes the mechanism more complicated and creates information leakage between the mechanism and the particular use case: future modifications to the use case are likely to require changes to the underlying mechanism as well.
In this case, the selection and cursor were not closely enough related to combine them. When the code was revised to separate the selection and the cursor, both the usage and the implementation became simpler. Separate objects provided a simpler interface than a combined object from which selection and cursor information had to be extracted. The cursor implementation also got simpler because the cursor position was represented directly, rather than indirectly through a selection and a boolean. In fact, in the revised version no special classes were used for either the selection or the cursor. Instead, a new Position class was introduced to represent a location in the file (a line number and character within line). The selection was represented with two Positions and the cursor with one. Positions also found other uses in the project. This example also demonstrates the benefits of a lower-level but more general-purpose interface, which were discussed in Chapter 6.