Information Technology Reference
In-Depth Information
this are to be found in the requirement to bring certain parts of the business to
market early and competitively. From this mile stone backwards scheduling is
undertaken to calculate the date of putting into operation and therefore in the end
the start date of software development and other previous mile stone dates
(functional specifications, technical specifications). Since development capacity is
limited time pressure arises as a matter of course.
One possibility to gain more time is to avoid long pilot operations. Pilot oper-
ation in this context means the simulation of real operations on a test system, which
mirrors the real world including its complete data base either before going into real
operation or in parallel with it. Pilot operations until final clearance for going life do
not only require calendar time but will bind development, quality and especially
end user resources extensively.
It is quality management that can influence the optimisation of time and
resources by taking care that nearly error free software can go into pilot operation
in as much as the latter is still deemed necessary. Ideally the acceptance process
itself should be sufficient making pilot operations completely superfluous.
3.2.2 Avoidance of Production Breakdown
With or without pilot operation the objective is to avert negative consequences after
putting software into operation, which has not been tested sufficiently. This sounds
quite obvious. But total system breakdown is not always necessary to let daily
business grind to a halt. Inefficient control of screen sequences, systematic data
errors from updates via interfaces or the failure of a critical algorithm suffice to
block the work schedule of users. And exactly these kinds of errors or inefficiencies
do not show up during development tests.
A careful selection of test beds as well as a thorough and comprehensive design
of test scripts allows for a maximum number of variants for a business process to be
tested. As elsewhere: the point is that the effort to be invested has to be exerted in
any case. If it is not done at the early stages, costs will multiply later on.
3.2.3
Instant Resolution of Faults
The discussion regarding error handling shows clearly that error detection and
amendment loops are most cost-efficient and controllable during acceptance tests
by quality management. The identification as well as control of measures to be
taken resulting from error detection during production at a later stage will generate
significantly more communication overheads compared to error detection during
acceptance tests by project specialists, who have participated in drafting the
