Container (Control)

Based on a task analysis, allow a container to either support a single, browse, multiple, or extended selection model.

If you design a container so that the container as a whole can take focus, then the following should occur:

When a container as a whole can take focus, the following should occur:

When the container as a whole has focus, a transfer operation whose destination depends upon the focus (for example, Paste) transfers into the object represented by the container as a whole. If the elements being transferred are not objects, the application encapsulates them as objects where this is possible.

Design a container so that the container as a whole can take focus if support for transferring objects to the container is required, unless the user can switch to a spatial view by using a graphics cursor.

In a spatial view, either use an element cursor or a graphics cursor.

In a linear view, use an element cursor.

In a linear view, when the active cursor is on an object, pressing ↓ (except on the last object) moves it to the object in the row below.

In a linear view, when the active cursor is on an object, pressing ↑ (except on the first object) moves it to the object in the row above.

In a linear view that is neither detailed nor hierarchical, pressing → has the same effect as ↓, and ← has the same effect as ↑. For bidirectional and vertical language support, see Chapter 11.

In a linear view, to represent an object use a small icon with the text label adjacent to the image if both are present.

In a linear view, do not allow the icon's text label to be cursored separately from the icon image. If you want to allow the user to change the label text, support direct editing techniques, property-editing dialogs, or both.

In a linear view, only the text label or the image of an icon needs to be shown. However, whichever one is present uniquely identifies the object.

Support the range technique when multiple elements can be selected in a container that uses a linear view.

In a linear detailed view, either all detail elements take focus or they all do not take focus. If they do not take focus, and you want to allow the user to change the details, support direct editing techniques, property-editing dialogs, or both. Detail elements are not selectable, even when they can take focus.

When multiple elements in a row of a nonhierarchical linear view can be cursored (because detail elements can take focus in addition to the icon in each row), follow the standard rules for directional navigation; that is, ← and → traverse through rows, and ↑ and ↓ traverse through columns, with the objects treated as a single column.

In normal mode, if keyboard navigation is used to move focus to a detail element, deselect all objects in the container.

If the user clicks Ctrl SELECT on a detail element in the container that can take focus, move the focus to that element without affecting the selection state of the objects within the container or having any other effect.

If the user clicks SELECT on a detail element in the container that can take focus, move the focus to that element. If select mode is used within the container, all objects are deselected as well.

In a linear hierarchical view, each object that contains subobjects should be displayed with a display control button to its left. The button should contain a value choice that may be toggled between an “expanded” and a “contracted” state and should not take focus.

When an object's display control button is in the contracted state, the subobjects of the object should not be shown; in the contracted state, they should be shown. All of an object's subobjects should be laid out vertically immediately below the object, but indented to the right (only the object column — not the detail columns). If any subobject itself contains subobjects, it should also be shown with a display control button that indicates whether its subobjects should be displayed. For information on bidirectional and vertical language support, see Chapter 11.

When focus is on an object with a display control button and the user presses Ctrl ←, set the button to its expanded state. If the user presses Ctrl →, set the button to its contracted state. For information on bidirectional and vertical language support, see Chapter 11.

If the user clicks the SELECT button on a display control button, toggle its state and move focus to its corresponding object icon.

Toggling a display control button (either with a standard keyboard or mouse) has the following effect: