Mapping data objects to interaction objects

Data is the basic ingredient of information. In the previous blogs, I outlined those parts of the steering model that are subject to variation. Those information-shaped elements are extracted from various internal and external data sources. The internal structure (meta data model) of the data source can give a large amount of extra information that is useful in designing effective interfaces in particular in rendering graphs.

Enlarge the system with meta data

In the following blogs, I shall enlarge which meta data attributes are useful for the design of the user interface. First, I cover the subject of how to select the most appropriate interface control for a particular object attribute. Secondly, I will inspect those properties that make it possible to automatically create the best presentation for a specific relation within a data set. Managers do not only retrieve information from the system, but are also responsible for formulating policy and strategy, planning figures and setting norms. Consequently, figures regarding the managers planning responsibilities should be editable. Other information should not.

It provides guidance for control selection

Meta data provides guidance in control selection that is essential if the flexibility of the graphical user interface is to be fully exploited. The control selection should be based on a classification of the data type of the attribute. Based on these values the system should apply a different set of rules. The mechanism for numeric (integer, real) data types is based on the following variables (Baar, et all, 1992):

• content – should the data be editable or not;
• limits – can the upper and lower bounds of values be determined in advance;
• range – the relative size of the possible set of values;
• precision – is the user more concerned with an approximate or accurate value .

Approximate of nature

We have already discovered that planning key performance indicators are approximate of nature. Therefore, planning figures can be set to a low precision. The same holds true for the content variable. Planning figures are editable and current information not.

Enumerated attributes are often used to fill dimensions with categories. Their optimal corresponding selection control can be determined by considering the following variables (Baar, et all, 1992):

• number of items – few (9 or few choices) or many (10 or more) items;
• set size – static (number of items fixed at runtime) or dynamic (varies at runtime);
• label – the length of the label for the longest choice in the enumeration;
• min/max – the minimum and maximum number of simultaneous choices.

The classification that indicates the number of items can be derived simply by counting the amount of categories (and considering a possible applied filter) of a dimension.

Selection rules for strings

Selection rules for strings are straightforward. They consider only the variables content and length. A boolean data object can be simply mapped to the interface object check box. For an overview of the different decision tables, I refer to Baar et. all., 1992.

Consider an abstract domain model

The rules used for control selection only consider the abstract domain model of a given domain in order to derive controls for editing or displaying data. I did not discover here directives that can draw conclusions from the domain model into concrete interface action controls such as buttons or menus. In the next paragraph, I outline the variations hidden in the data that direct the choice for effective graphical presentations.