Estimating the Testing effort
To gain an appreciation of the resources that
must be assigned to Acceptance Testing, there
are several ways to estimate the effort required
for individual stages of the Test process.
In
addition, web-based systems present their own
‘peculiarities’ and are assessed using separate

The Acceptance Testing Kit – Part 1
21
Function points
are a technology-
independent method of estimating
the amount of work associated with
developing a system.
criteria.
Any and every project should be subjected to some degree of Acceptance
Testing.
Preparing the Acceptance Test Plan
In general, the Acceptance Test Plan preparation phase is usually about 2% of the
overall effort.
As an example, a system that requires 4000 hours of development will
require in the order of 80 hours of resources dedicated to Test planning.
Conducting the Acceptance Test
The expected duration for an Acceptance Test is also derived from the system’s value
in function points.
For each 1000 function points, usually allow one calendar month
of Acceptance Testing.
Precise duration is more evident once the detailed
Plan and “bottom up” estimates have been prepared,
and in light of the number of participants and their
availability.
Testing Web-Based Systems
When the system is ‘web-based’ or has web components, testing effort should be
increased to allow for full useability testing and for testing against different browser
versions and platforms.
Web pages are subject to many types of web interfaces and
require more extensive useability testing than standard structured and controlled
graphical
user
interface
screens.
Web-based
systems
have
a
number
of
characteristics that necessitate additional care during testing:
►
There is less (or no) control over the environment in which the system will run.
Generally, the environment for an Intranet based system can be controlled to
some extent, whereas for a system on the World Wide Web, there is no control;
►
The system may need to run using different browsers and under a number of
different versions;
►
There is little opportunity to train end-users and there actions and logic are not
easily predicted;
►
There is no capability of controlling the sequence of events that an end-user
may invoke.
Consequently, testing of web pages is often less structured, allowing the testers free
reign in their (test) use of the system, possibly with an observer or recording
mechanism to monitor events.
Test Plan preparation will need to allow for this.

The Acceptance Testing Kit – Part 1
22
Why doesn’t our IM Branch do the Testing?
Before any system is handed over to users for Acceptance Testing, the system
suppliers and/or developers have tested it thoroughly at various levels.
These tests
are usually technically-focused and only one part of the Acceptance Testing process.


You've reached the end of your free preview.
Want to read all 96 pages?
- Fall '18
- acceptance testing, Acceptance Test Manager