From the information pile to social, knowledge exchanging bots
Traditional CMS greatly simplify the way data is accessed but the do not represent a higher level of abstraction than the underlying DBMS
- Possibilities for user customization are limited to dumb filters
- Recent smarter systems suggest documents basing on similarity with other users. Same technology as MovieLens (implemented in Drupal)
- The semantic web paradigm can really make the difference
- Relativistic data storage: the traditional separation between data and metadata
(structure) disappears. Abstractions such as meanings and other entities
external to the data universe take their functional role. As a consequence data
can not only be filtered and ordered in different way, but the same data can
represent different concepts. The computer can represent different views of the
world, and best of all, allow "learning" even between incompatible world
views. (This corresponds to the real world experience that communication
works even between people with contradictory ontological views)
Late interpreting, what we try to capture and actually store are brute facts/sensations. Interpretation and reasoning (inference) takes place when content is accessed, thus the interests and believes of the user can be taken into account. - CMS communicate with each other through their "B2B" interfaces or an enhanced Blogger API. Enabled with semantic web technologies they can learn from each other. User ratings put them on track to accumulate the "right" knowledge (value system).
- P2P finally efficient, because as in the real world, you ask those who might know. Social discovery networks already exists (alpine, neurogrid) but without knowledge representation they work only in communities that are enough homogeneous, at least in what terminology is concerned.
Tutorial
- Background (about the presentation outlined above)
- Assembling of a trivial RDF based CMS with jena
- Have multiple instances exchanging content
- Have exchange of content with non CMS (annotation system)
- Discussion of available open source projects such as neurogrid, samizdat
- Available standards /need for specification
About Reto Bachmann-Gmuer
I was born the 1976/2/1, studied philosophy, informatics and sociology at the universities of
Zurich and Fribourg (Switzerland). I work as external consultant for Java and knowledge
management. I engaged in various open source projects such the indymedia CMS "mir" . The
most recent project I initiated is WYMIWYG (what you mean is what you get) a project that
aims to build software with the semantic web and P2P paradigms to change the way reality is
(re)produced.
reto@gmuer.ch