Powl – Semantic Web Development Plattform

Overview

The broad application of ontologies as shared terminological knowledge representations is one of the main strategies of the Semantic Web paradigm. With RDF/S (Resource Description Framework) and OWL (Web Ontology Language) there exist W3C standards for defining web enabled ontologies which fit in the semantic layering of web languages.

Although there is a bunch of OWL ontology editing and management solutions available, some of them are complicated to deploy or handle, some do not support strategies for collaborative, distributed development of ontologies, some are not Open Source or not available for the most distributed web technologies.

Since PHP is by far the most distributed web development technology (as regularly confirmed by Netcraft), the Semantic Web paradigm will probably only be successful in a large scale if there are applications and tools available tightly interacting with this language. The aim of the Powl project is thus to deliver a PHP and web-based ontology edititing and management solution to the Open Soure community.

Architecture

Powl consists of the following components:

• RAP – RDF API for PHP an independent project by Chris Bizer, Radoslaw Oldakowski and others (thanks here for their great work)
• RDFS and OWL API for PHP – integral part of Powl
• Powl Webapplication framework:
  • Tabplugins, as Models, RDF-Triples, Classes, Properties, Instances – future releases may include e.g. export or visualisation plugins
  • Widgetplugins for data editing – widgets for comfortable editing of WYSIWIG-HTML, Dates, Tree-Selects are planned

Powl requires a PHP (greater 5.0.1) enabled webserver and a database backend for storing model data (My SQL, Postgres, Oracle and others – see ADODB Documentation for supported RDBMS).

Please have a look at the planned Powl architecture description for details.

Features

• Powl supports viewing, editing of RDFS/OWL ontologies of arbitrary size. It is even quite fast with the largest available models (e.g. NCI Cancer Ontology containing about 28,000 classes).
• Sophisticated widgets for data editing such as widgets for editing HTML in a WYSIWIG manner or for dates are integrated.
• Questioning the knowledge base. Powl currently offers an RDQL query builder as well as a full-text search for literals and resources.
• Plugin concept. Powl is easy extensible, unfortunately still laking exhaustive documentation on this – please have a look at the source code and pester the developers. :-)
• Powerful object oriented API. All functionality is accessible by a clean application programming interface.
• Authentification scheme. Fine grained exposition of features and model data: Privileges (view, edit) for users and groups are planned to be assigned to Models, Classes and Properties.
• Versioning. All edits of a knowledge base may be logged and rolled back (depending on time, user and edit action).
• Powl is fast. Models are stored in database tables and only those parts of the model are loaded into main memory which are actually needed. Thus Powl is scalable and fast.
• Multi language support. Powl comes with English and German translations of the user interface. If you would like to provide a translation please contact us!

You miss a feature? Please submit a feature request at the Sourceforge Project Page.