Databases Reference
In-Depth Information
Figure 1: Reference model of the Web structure construction task
(1) Program Development Approach
This approach is the most primitive one. Many projects to develop practical systems
have taken this approach so far. In this approach, both the data acquisition and
the data layout are performed in hand-coded application programs (Figure 2(a)).
Those programs are written in general-purpose programming and scripting
languages such as perl and Java. They usually include access requests to the
information sources and database queries. They have to generate HTML documents
as their output. Those tailor-made application programs are invoked through
frameworks such as CGI and Servlet
44)
, and Web Server APIs such as ISAPI
33)
and
NSAPI
38)
. When invoked through CGI, programs are associated with URLs.
Figure 2: Three major approaches
(2) Template-based Approach
Since the Web structure construction problem has been recognized as an important
issue for application development, more sophisticated approaches have emerged.
They include the template-based approach (Figure 2(b)). In this approach, templates
are used to give data layout specification. Typically, the templates are based on
HTML. Data acquisition specification is given as queries and scripts embedded
into those templates. The approach is similar to the program development approach
in that the data acquisition and the data layout are specified together. However, it
has a number of advantages over the program development approach.
In this approach, users need not write programs from scratch. In particular, con
struction of templates is far easier compared with writing program codes to generate
