Our approach is a kind of gray box testing tool, so we could
make use of the system database and employ the data kept in it
to fill forms. One of the ontologies used in this project is the
system ontology, which is learnt from the system database.
Another ontology which is employed is built as follows: More
than a hundred forms are downloaded from different web sites.
Afterward, their fields are categorized and synonym words are
eliminated using the WordNet
1
ontology. The result is about
70 words remained, which are utilized for constructing the
ontology. We refer to this ontology as general ontology.
General ontology is equipped with defining accurate
constraints. We use general ontology and its constraints to
produce data which are not kept in system database.
The resulting model is even computable in the presence of
high extreme dynamism. It is also obtained with a high degree
of automation in the dynamic behavior of web applications. If
the inputs used in this generation could cover all relevant
behaviors of the web application, the model generation is
completed. This is an intrinsic limitation of all dynamic
analyses. The tool we propose to support the recovery of
structural model from client side analyzing is exhaustively
described in the following section.