Principle 17 - Applications Ensuring Data
Quality
Applications must help ensure valid, consistent, and secure data.
Note
As used here, the terms "valid" and "consistent" refer to data which satisfies
a set of business rules. Documentation defines what is valid and consistent for
a data element or set of related elements. Data which is valid may or may not
be accurate (that is, "correct" or "factual"). Applications cannot verify the
accuracy of the data.
 Rationale
It is unreasonable to place the full burden of data validity and consistency
on manual operating procedures or the quality of the training provided for the
users of the application. Applications are the vehicles for creating and using
data and are therefore uniquely positioned to help ensure the validity and consistency
of the data.
Implications
- The application's creation and use of data must conform to the rules for
that data as specified in the University data architecture.
- The application is only one of several possible vehicles for verifying data
validity (database tools such as active dictionaries are another). These vehicles
must work together to ensure optimum results.
- Similarly, the application is only one of several possible vehicles that
can be used to secure data and these tools must work together to ensure optimum
results. Security restrictions implemented at the application level must follow
standards for security and confidentiality but should not inhibit flexibility
or ease of use.
- Wherever possible, common data validation routines should be developed as
reusable components are shared by multiple applications.
- It is not always possible nor desirable, given cost/benefit or operational
considerations, for an application to verify data completely or immediately.
The application's role in validating data lessens the burden but does not lessen
the responsibility of individual users for ensuring overall data quality. Associated
training and support for users are ongoing requirements.
|
