$Revision: 1.4 $ $Date: 2002/05/29 12:24:51 $


SemWalker is a Semantic Web Browser which runs as a server-side application so users don't need to install any software. It embodies one concept of what the Semantic Web is: a shared database of property/value pairs associated with objects. This database is global, partially available, partially trustworthy, sometimes modifiable, and even sometimes contradictory, just like the rest of the web. SemWalker lets you see and modify information in this database.


SemWalker is open source. It's purpose is initially demonstrative, but we hope it will grow into something very practical.


@@ prereqs

pointer into ?

Use it On-Line

Like any browser, SemWalker needs a place to start. Normally it starts with an object identifier, via a hypertext link, as with these links to: Sandro and Semantic Web Technology.

It also has a Query Form, for when your starting address is in plain text. It should also allow page and pattern queries.


SemWalker is build to use a content-programmable proxying triplestore interface, which allows it to concentrate on human user-interface issues. The data operations it performs are available to any other program using a CPPT implementation.


The primary use case is this: the user provides an object identifier (usually via a URL query parameter) and SemWalker displays information about the object. There are many possibilities for (1) where the data comes from, (2) how it gets transformed before being displayed, and (3) what of it is displayed.

There are additional modes for displaying a collection of objects, a collection of statements. There are (or will be) also ways to modify the data in the system.

The configuration of SemWalker itself is available for viewing and modification as with other data.

Here's a more detailed list of features, which may turn into milestones and checklists and stuff.

Display properties/values for an object, using data fetched from the object's URI

When displaying a propery or value, use rdfs:labels

Hourglass display

Give all the links in on the left, all the links out on the right, so it's like looking at one node, with all its links, from a circles-and-arrows diagram.

Session Controls

Track sessions to allow for customization and proxy authorization.

Add new data

Remove data

Inference (monotonic, processing, gathering)

View old data

View with multiple trust levels

Access to protected sources

Display search results -- list of objects

Display search results -- list of triples

Display legacy representations

Sort and control truncation

Redirection to alternative SSSWBs

Sandro Hawke
First: 2002-03-29; This: $Id: Overview.html,v 1.4 2002/05/29 12:24:51 sandro Exp $