Reducing Hierarchy Levels

Various options are available for "slimming down" hierarchies, and they differ in their usage and simplicity of implementation. The common starting point of all these techniques is moving the upper-most hierarchy levels to a type of "filter" or "view" mechanism. These variants are called tabstrip, filter and shuffler.

TabStrip

Figure 1: Converting nodes to tabstrips

A tabstrip allows the user to switch between different views of a dataset. Due to the relatively large space requirements, only a small number of nodes (3-5) can be modeled.

Summary: Tab strips

Display different views
Display a small number of categories

Filter

A filter reduces the amount of displayed data by setting a criterion. Different design options are available for entering the criterion. The example in the following figure shows the conversion of hierarchy level "Flight number" to a drop-down list box.

Figure 2: Filter for reducing the number of hierarchy levels

The advantages of the filter solution are its relatively low space requirements and its ability to model a relatively large number of nodes. However, not all the "views" are visible at the same time, and the user has to perform an additional interaction to change the view.

Summary: Filter

Suitable for modeling nodes
Reduces a medium-sized number of categories
Data is filtered and displayed as views

Shuffler

The shuffler concept extends the filtering of data described above with two additional aspects. First of all, a shuffler usually consists of a collection of queries, for example, a combination of order numbers created during a specific time interval. Secondly, the variable components of the query are embedded in a semantic relationship - that is, a single sentence. For example, this sentence could say "Show me all the order numbers that were created in the last year". Theoretically, each variable component enables you to roll out one hierarchy level in a structure.

Figure 3: Example of a shuffler

The figure shows an example of how a shuffler can be used to reduce a three-level hierarchy (object type, processing type, and processing interval) to a simple list.

Summary: Shuffler

Suitable for modeling complex dependencies
Formulates the filter criteria in "natural" language
May be complex to program

top