Managing Information Technology

Chapter 4 • The Data Resource 103

In this figure the processing of data into information
is viewed like the processing of the raw and component
material resources into final products in a manufacturing
company. Raw data are captured or received, inspected for
quality, and stored in the raw material, or operational, ware-
house. Data in storage are used in the production of any
authorized information product (e.g., report). Data are
retrieved from the warehouse when needed but, unlike raw
materials, are not consumed when used. As operational data
become obsolete (e.g., product price increase), they are
replaced with new data. Somewhat different from our manu-
facturing analogy, in order to understand trends and patterns,
snapshots of data are periodically transferred to a long-term
data warehouse. The long-term data warehouse allows
historical transactions to be understood within the context of
the characteristics of customers, products, facilities, and the
like at the time of the transactions, not just the current
situation. Data are transferred from operational and histori-
cal sources to other parts of the organization or other
organizations when authorized. As data are processed into
information (e.g., sales forecast), this information is added
to the warehouse, similar to the entry of products into
finished goods storage. All operations and work centers use
the raw material warehouse to produce information products
(e.g., reports), but individual work centers (applications)
may have their own work-in-process inventory of data and
receive a few kinds of data that are not shared among other
applications. Thus, data are cataloged, managed, and, at
least conceptually, stored centrally, where they can be kept
safe and uncontaminated for use throughout the business.
The central point of Figure 4.4 is that data and applica-
tions software must be managed as separate entities. When
treated separately, data are not locked inside applications,
where their meaning and structure are hidden from other
applications that also require these data.

APPLICATION SOFTWARE CAN BE CLASSIFIED BY HOW
THEY TREAT DATA The concept of application inde-
pendence suggests that different data processing applica-
tions can be classified into three groups, based upon their
role in managing data: data capture, data transfer, and data
analysis and presentation, as shown in Figure 4.4.
The process of transforming data into information
useful for transaction management or higher-level decision
making includes these steps:

1.Data capture Data capture applications gather data

and populate the database. They store and maintain

data in the data pyramid of Figure 4.3. Ideally, each

datum is captured once and fully tested for accuracy

and completeness. Responsibility for ensuring the

quality of data capture systems might be distributed

across the organization. Localized data capture

applications are developed for data with an isolated

use or data for which coordination across units is not

required. Still, because localized data might eventu-

ally be useful somewhere else in the organization

(and they must then be consistent across sites), an

inventory of data elements (in a metadata repository)

must be maintained of all database contents.

2.Data transfer Data transfer and integration appli-

cationsmove data from one database to another or

otherwise bring together data from various databases

to meet some processing need. These applications are

often called bridges or interfaces because they connect

related databases. Once raw data are captured, they

might be extracted, transformed for new purposes, and

loaded into various databases where they are stored for

specific purposes. For example, customer order data

might be stored in multiple subject or target area

databases supporting production scheduling, billing,

and customer service. Also, this kind of application

extracts and summarizes data as well as distributes

copies of original data. Ideally, this transfer would be

event triggered; that is, if new basic data are captured

or changed in value, messages are sent as needed to all

other databases that build on these data to alert these

databases that changes have occurred.

3.Data analysis and presentation Data analysis and

presentation applicationsprovide data and informa-

tion to authorized persons. Data might be summarized,

compared to history, reformulated into graphs, or

inserted into documents being developed using a word

processor. Data might be input to a decision support

system or executive information system (to be dis-

cussed in Chapter 6). Data analysis and presentation

applications can draw upon any and all data from data-

bases the business manager receiving the presentation

is authorized to see. Data and the way they are present-

ed should be independent, and those who determine the

format for presentation should not necessarily control

the location and format for capture and storage of data.

APPLICATION SOFTWARE SHOULD BE CONSIDERED

DISPOSABLE A significant result of application inde-

pendence is the creation of disposable applications.In

many organizations, older systems cannot be eliminated or

easily rewritten because applications and data are so

intertwined. When the presentation capabilities of an

application system become obsolete, but the application

also maintains data that are essential to the business, an

inefficient system might have to be kept alive only for its

data access capabilities. With application independence, a