Information Technology Reference
In-Depth Information
avoided such “mistakes,” the screen or report would become clearer. In
this vein, Tufte mentions magicians. One of the key ways they minimize
the transfer of information to the audience is through the use of distraction
— something is swirling in the left hand while the right hand switches
the coin. Reversing this logic, if one wants to maximize the transfer of
information to the user, one should minimize the distraction — that is
why Web pages with revolving globes can be distracting and are consid-
ered bad design.
Determining the Many Uses
In one sense, data is neutral. It stands ready to serve many (information)
masters. From the same data, different information can be drawn. One
can use the end-of-day inventory to calculate daily sales, as well as make
replenishment decisions. One can use the date of a person's employment
to assign him or her to a particular performance review cycle or to calculate
retirement benefits. The flow between data and information is “one to
many.” In such a situation, data design must anticipate the many uses.
This is done through system requirements. This is also what makes
enterprise data models difficult to design and implement.
Organizing Data
The conversion of data to information involves two basic activities: (1)
organizing and (2) processing. Organizing data is reflected in the data
models underlying the application. Applications are built on these data
models. If both the data model and the application are designed at the
same time, one would expect the fit to be there. Issues arise when an
application is expected to use an existing data model “belonging” to
another application. The data model may have to be modified. Any change
to the data model can be expensive as “data serves many masters.” There
can be side effects and ripple effects.
Data models deal with data related to entities and the relationships
between entities; for example, a customer can have many orders. Adding
a new piece of data (e.g., cellphone number) is easier than adding a new
concept related to customers such as “national customers” and “interna-
tional customers,” because the latter introduces new semantics within the
data model and the applications residing above. Such changes are often
required because the
(refer to Chapter 2) involved is not recognized
properly. What you thought was a single entity called “customer” turns
out to be two differing entities.
variety
