2.5 - Components of the WebLinker System

WebLinker consists of two distinct subsystems:

2.5.1 - WebLinker Traversal Robot

In order to modify external references within the infostructure and to create tables of internal references, we need to be able to traverse an entire infostructure parsing the HTML files to extract the relevant information.

MOMSpider is used as the body of the traversal robot. This has hooks into the WebLinker routines, so that we can parse the contents of every HTML node that MOMSpider visits.

The robot is used to do two things:

Generate the table of internal references for an infostructure.
Update logical links to external infostructures to include Logical Resource Names.

2.5.2 - WebLinker Server Subsystem

In order to do the LRN to URL conversion dynamically, one script is needed on the server storing the infostructure. This is called as a Common Gateway Interface (CGI) [5] script. On the server, we designate a virtual path that will represent the location of infostructures. In this we place the mappings table for all the infostructures, as shown in Figure 3.

Figure 3: The structure of an infostructure on the filing system. The lrn directory contains a file giving the LRN to URL mappings for each infostructure (there are two here, FARMING and Fish), as well as a BASE element, giving the physical location of the infostructure on the server.

When a HTTP request referring to an infostructure is received, the script is called. It then works out which document on the server has been requested, and returns the contents of that file.

One extra feature provided by the use of a script is that it allows us to return extra information to the client. We define a special LRN as follows:

For an infostructure FARMING on the server www.agri.org, we define the

LRN http://www.agri.org/lrn/FARMING?frags

to return the complete list of valid fragment identifiers for that infostructure in text/plain format. This allows the WebLinker robot to use such a LRN to retrieve the fragment identifier mapping table for a remote infostructure easily. Also, the

LRN http://www.agri.org/lrn/FARMING?List

returns a formatted HTML document giving the list of fragment identifiers and a brief description of what they are. This can automatically be extracted from the HTML nodes in the infostructure: Using the HTML fragment from Section 2.3.3 as an example, the following descriptions are generated automatically, one for the fragment Pastimes and one for the file farming_33.html itself:

Pastimes                    Pastimes of sheep farmers
farming_33.html             Bolivian Sheep Farmers

WebLinker@ptsun00.cern.ch - 16 SEP 94

[Next] [Previous] [Up] [Top] [Contents] [Search]

Generated with WebMaker