Information Technology Reference
In-Depth Information
There is a need for a framework for measuring the quality of converted databases.
The following criteria are derived from the requirements of software engineering
and database technology:
• Integrity—Only syntactically and semantically correct data should be stored in
databases to enforce domain integrity. Referential integrity is another type of se-
mantic integrity such that data cannot exist or be modified unless some precursor
data values exist or some actions are taken.
• Trace-ability—Agooddatabasedesignshouldsupporttrace-abilityfromthe
requirements down to the physical design stage back through documentation.
Trace-ability is necessary for different phases of database development. Simpli-
fication and overload errors can occur in any phase and will affect the degree of
trace-ability.
• Consistency—Indistributeddatabasesystems,dataareoftenreplicatedtoim-
prove performance and availability. All copies of the same logical data item must
agree on exactly one “current value” for the data item. All users within the envi-
ronment should have a uniform view of the system. If the data is inconsistent, the
users cannot share the same information. It is particularly important for parallel
applications that partition data into different parts to increase their processing
speed. If the partitions are stored in different sites, consistency is a key factor to
ensure correctness of the application.
• Correctness—Adatabaseiscorrectifitcorrectlydescribestheexternalobjects
and processes that it is intended to model. They use a set of static constraints on
objects and their attributes, and a set of dynamic constraints on how objects can
interact and evolve. A database is said to be syntactically correct if the concepts
are properly defined in the schema at each stage; it is said to be semantically cor-
rect if the concepts are used according to their definition at each stage.
• Completeness—Adatabaseschemacanbedefinedascompletewhentheschema
represents all relevant features of the application domain. Two major principles
can be used to check completeness of the design: (1) checking all the require-
ments of the application domain and ensuring that each of them is represented
somewhere in the final system; (2) checking to see whether each concept is men-
tioned in the requirements.
• Efficiency—Adatabaseschemacanberegardedasanefficientdesignifthe
schema (1) can support any processes on the component schema; (2) provides
both timely and accurate data access for a given set of queries and transactions.
Information technologists have moved from data processing to information pro-
cessing and are now moving into the field of knowledge processing. The new term
EDS has emerged to refer to an important area in this field. An EDS is a system that
results from the integration of ESs and DBMS technology.
Consider the following problem taken from a real application: a personnel man-
ager must find the best person for a particular job, or the best group of people for
a particular project (i.e., a project that includes different types of jobs) by con-
sidering the total departmental manpower. A common way to solve this problem
is to send the employee information and the job vacancy information to a human
Search WWH ::




Custom Search