Creating applications for the Web is substantially different
from the rich-client approach. It implies that most of the work is
done on a remote server which sends ready-to-use images and scripts
to a Web client (usually an Internet browser). There are many variants
but the main idea is to receive and process HTTP requests on a server
and, in return, to produce the right image to be sent back to the
browser. Even if there are many architectural differences, the following
aspects can be shared between rich and thin clients:
Styling: responsible for transforming
data into graphics, the styling rules can be the same for rich and
thin-client applications. For example, CSS files, symbols and Designer
projects can be used for both scenarios.
Data model and connection with a data
source: are the same and can benefit from the same code for both deployment
strategies.
Graphical content: usually displayed
in the background of a panel or as parts of a dynamic symbol; graphical
objects can be reused in both cases.
The main differences reside in the way the application
behaves and how it is packaged. Thin clients are usually driven by
transactions and only partially support rich interactions. JViews
Diagrammer thin clients are based on the JavaServerâ„¢
Faces
(JSF) and DHTML technologies. Web pages containing visual components
are created as (JSP) pages. The JSF-based components (also known as
JViews Diagrammer Faces) provide the following services:
Display graphical content managed by
Diagrammer classes and APIs. Data connection and styling are preserved.
Provide interactive views and overviews.
Manage local interactions, such as pan,
zoom, selection, pop-up menus.
Minimize round-trips when possible.
Extensions can be written for extending
both server-side and client-side components.
The use of these JSF and DHTML components is described
in the section Building Web applications of
the documentation. This section also describes how to deploy resources
for your Web applications.