Building the Data Warehouse

Скачать в pdf «Building the Data Warehouse»

■■ Breaking a long-running program into a series of shorter programs ■■ Requiring a program to access a smaller amount of data ■■ Shortening the time a unit of data is locked ■■ Changing a lock from update potential to access only

PARAMETERS OF SUCCESS: Algorithmic specification is done correctly when all details needed for specification are identified, and only details needed for specification are identified, one step at a time, including all possibilities. In addition, conformance to the standard work unit (SWU) is ensured here.


PRECEDING ACTIVITY: Algorithmic specification; physical database design; data store design.

TIME ESTIMATE: Varies (see previous activity).

NORMALLY EXECUTED ONCE OR MULTIPLE TIMES: Once per activity (at the primitive level).

The algorithms and program specifications are further refined into pseudocode. The designer ensures that all needed data for processing is available. All variables used in calculations, transformations, updates, and so on are identified here. Any loose ends are identified. Performance at the design level is considered here. The output of this activity is coding specifications ready for actual code.

PARAMETERS OF SUCCESS: The final pass at programming specification includes the following:

■■ Completeness ■■ Order of execution ■■ All required cases

■■ All contingencies, including error handling and exception conditions ■■ Structure of coding




TIME ESTIMATE: Depends, from one day per activity to two weeks per activity. NORMALLY EXECUTED ONCE OR MULTIPLE TIMES: Once per activity. DELIVERABLE: Source code.

The pseudocode is translated into source code. If the data has been designed properly, and if the specification of pseudocode has been thorough, this step goes smoothly. The output of this step is source code.

Скачать в pdf «Building the Data Warehouse»