Information Technology Reference
In-Depth Information
...
Automation Framework
Significant development has happened over the last 5 years in the field of automa-
tion of testing. Before 1998, when record and playback features were unavailable in
commercial, off-the-shelf automation software, testing professionals used to create
testing scripts manually. It was a very resource-intensive task. Often almost the
same amount of effort was required to test a component as was required to develop
it. By 1998 Mercury Interactive launched their record and playback testing tools.
This made automating testing less effort intensive. But still there was a lot of effort
required in maintaining these scripts, as any slight changes in screen or data would
cause changes in the test script. The other problem with these scripts was that for
each test case, there was a script generated by the tool. So the same object with
test data was repeated several times in the script. Any change in the object would
require changing its properties at all of these places. This approach resulted in inef-
ficient utilization of resources.
So some new techniques started evolving after 2001. All of them required the
establishment of a framework that can be utilized when script is being written.
Some of them include a data-driven automation framework, a keyword-driven
automation framework, a hybrid framework, a model framework, and so on. The
common thread among all these frameworks is reuse and automation of testing as
much as possible. Traditionally it has been established that if a test case has to be
executed more than 10 times, then it makes sense to automate it, as this much effort
in executing the test case manually becomes equal to the effort in automating the
test case. If the test case has to run more times than this, then the test case execu-
tion effort will be much less than the automation effort and will keep on becom-
ing less afterward. Nowadays the approach is toward automating the entire testing
process from the very beginning and no manual testing is done at all.
... Test Data Management
Test data preparation, maintenance, and management are among the most impor-
tant activities for software test management.
5.8.1 Data Maintenance in Production Environment
In a production environment, how can you create and maintain test data? You have
live data that is being used by end users, and any erroneous, duplicate, or unrelated
data may wreak havoc on the production system. You are not supposed to tamper
with end user data as well. So you cannot use end user data for testing production
environments. So how you are going to keep testing production environments?
Well, you can manage and maintain test data in your production environ-
ments by creating a complete set of test data that can be easily identified from
Search WWH ::




Custom Search