SemanticsSortKey class

Base class for all sort keys for Semantics accessibility traversal order sorting.

Only keys of the same type and having matching names are compared. If a list of sibling SemanticsNodes contains keys that are not comparable with each other the list is first sorted using the default sorting algorithm. Then the nodes are broken down into groups by moving comparable nodes towards the earliest node in the group. Finally each group is sorted by sort key and the resulting list is made by concatenating the sorted groups back.

For example, let's take nodes (C, D, B, E, A, F). Let's assign node A key 1, node B key 2, node C key 3. Let's also assume that the default sort order leaves the original list intact. Because nodes A, B, and C, have comparable sort key, they will form a group by pulling all nodes towards the earliest node, which is C. The result is group (C, B, A). The remaining nodes D, E, F, form a second group with sort key being null. The first group is sorted using their sort keys becoming (A, B, C). The second group is left as is because it does not specify sort keys. Then we concatenate the two groups - (A, B, C) and (D, E, F) - into the final (A, B, C, D, E, F).

Because of the complexity introduced by incomparable sort keys among sibling nodes, it is recommended to either use comparable keys for all nodes, or use null for all of them, leaving the sort order to the default algorithm.

See Also:

• SemanticsSortOrder which manages a list of sort keys.
• OrdinalSortKey for a sort key that sorts using an ordinal.