Databases Reference
In-Depth Information
Problems with data can result from missing information and mismatches
between the old model (often only implicit) and the new model (usually ex-
plicitly documented). Problems also result if the conversion effort is start-
ed too late in the project and is under-resourced. The most common
sources of problems are data quality and incomplete data.
Costs and Benefits of Data Conversion
Before embarking on data conversion, the data conversion team should
decide whether data really needs to be converted and if it is feasible to
abandon the noncurrent data. Starting fresh is an option.
The customers may decide that the cost to preserve and correct old in-
formation exceeds the benefit expected. Often, they will want to preserve
old information, but may not have the resources to correct historical er-
rors. With a data warehouse project, it is given that the data will be con-
verted. Preservation of old information is critical.
The Cost of Not Converting
The DCT first should demonstrate the cost of permitting erroneous in-
formation into the new database. It is a decision to be made by user man-
agement.
In the long run, permitting erroneous data into the new application usu-
ally will be costly. The data conversion team should explain what the risks
are to justify the costs for robust programming and data error correction.
Costs of Converting
It is no easier to estimate the cost of a conversion effort than to estimate
the cost of any other development effort. The special considerations are
that there may be a great deal of manual intervention, and subsequently ex-
tra programming, to remedy data errors. A simple copy procedure usually
does not serve the organization's needs. If the early exploration of data
quality and robust design and programming for the conversion routines is
skimped on, IS generally will pay for it.
STEPS IN THE DATA CONVERSION PROCESS
In even the simplest IT systems development projects, the efforts of
many players must come together. At the managerial and employee levels,
certain users should be involved, in addition to the applications develop-
ment group, data administration, database administration, computer oper-
ations, and quality assurance. The responsibilities of the various groups
must be defined clearly.
In the simplest terms, data conversion involves the following steps:
Search WWH ::
Custom Search