Continuous capture mechanism

This section describes the basic mechanism used to associate continuous captures with an initial content capture.For a portlet that is defined with continuous capture, one or more URL patterns are associated with a Content Capture Language (CCL) command. The CCL command extracts features from the page that was reached by clicking a link that matches the URL pattern associated with the CCL command.

Web elements do not store static HTML, but are defined by a CCL statement that dynamically retrieves the content from its source. The CCL statement consists of:

The location of the Web page, and
The feature tag of the captured content. This feature tag is an electronic signature of the content and consists of a feature list that describes the content.

When Web Studio executes the CCL statement, the capture technology parses the Web page and returns the content that matches the feature tag. If an exact match cannot be determined, the content with the closest matching feature tag is returned. This allows Web Studio to provide the correct Web element even though the content, location, or formatting of the source may have changed.

Continuous capture defines a current “path,” which groups the available URL-CCL mappings. This path begins at the top level as “/”, and all URL matches at the top level are then labelled “/1”, “/2”, and so on. The tree path defines all continuous capture possibilities that exist for a portlet.

The continuous capture definition is a set of mappings for each level of capture that relates URL expressions with a CCL command. Each mapping describes how to extract the required content from a Web page that matches the URL expression.

When the portlet plays back in Portal Interface, users click a link on the initial page. The URL they click matches a top-level URL expression and a corresponding CCL command. This combination returns the page content. Users can then click links on the returned Web page, and using URL expressions defined for this level in the playback, find matching CCL commands that determine how that page is returned.

If there is no URL expression that matches the clicked URL, the entire page capture is returned.