Selenium was used to execute test scripts in this approach.
Selenium has the ability to give us useful feedback about
results of execution. It automatically produces verdicts. The
final column of table 1 shows the number of steps
automatically done by each tool and especially by the
proposed approach. The results indicate that automation done
by our approach is the greatest one.
The experiment on WebCalendar is done in 5 phases:
In the first phase, WebCalendar is set up. Then, several
users are identified to use WebCalendar software for
scheduling their times and managing their reminders and
events. All events entered into the system are associated with
one or more users. There is no limit to the number of users that
can use WebCalendar. Each user has a login and a password,
which are stored in the database. The database is named
‘intranet’.
In the next phase, we run MBTester to transform ‘intranet’
database to the OWL ontology format. All data stored in
database are transformed to instances in the ontology.
After the transformation process is completed, in the third
phase, MBTester is used to map ontology that is learnt from
‘intranet’ to general ontology. WordNet has been used to
generate synonymous and antonymous words. With the help
of Jena
4
library, we could use WordNet feasibility in Java.
In the fourth phase, MBTester is run to exploit structure of
WebCalendar, including static and dynamic aspects. The
structure is stored in the intermediate format.
In the last phase, all pages are navigated for detecting errors.
MBTester reads structure stored in database and start from
base URL, all hyperlinks included in this page are opened for
detecting broken links, and all forms which are included in the
page, are filled and submitted for discovering errors.
The results obtained by running MBTester on WebCalendar
application are discussed later. Filling form means generating
data for parameters needed. In fact, it was specified as a
second step shown in table 1. Table 2 indicates the degree of
automation supported by the proposed tool for filling each
separate form. The second column illustrates the number of
elements which are included in the form. MBTester tries to do
mapping between each text field on the form and concepts on
general ontology. The number of wrong mappings is stated in
third column. At the final column, the total degree of
automation supported in filling every form is demonstrated.
The last row specifies the total automation supported by
MBTester in the second step of manual testing process.
Table 2 identified the model coverage supported by the tool
to the depth 3. It is important to note that model coverage is
absolutely different from code coverage and the proposed tool
only support model coverage.
The results indicate that automatic support for the
verification and validation activities by MBTester has
increased when compared to related works mentioned in table
1. The only user’s interference is during the mapping between
two ontologies in which user should approve or correct the
tool’s suggestion.
4
Jena is a Java framework for building Semantic Web applications.
http://jena.sourceforge.net/