The existing, open Worldwide Web has been successful on a global scale because the cost of participation at a basic level is low and the individual benefit of participation is immediate, rising rapidly as more participants take part. The same cannot currently be said about semantic based systems because the cost of being precise about semantics for sophisticated components is prohibitively high and the cost of ensuring an individual, absolute semantics for a component rises rapidly as more participants take part. OpenKnowledge aims to break out of this deadlock by focusing on semantics related to interaction (which are acquired at low cost during participation) and using this to avoid dependency on a priori semantic agreement; instead making semantic commitments incrementally at run time. The "Open" in OpenKnowledge thus is significant in two senses: it assumes an open system, which anyone may join at any time; it assumes an openness to being joined, achieved through participation at low individual cost.

OpenKnowledge is a system which allows peers on an arbitrarily large peer-to-peer network to interactive productively with one another without any global agreements or pre-run-time knowledge of who to interact with or how interactions will proceed. Existing services can be translated into OpenKnowledge peers and can then make use of the full functionality of the system: this translation process already exists for WSDL services and can be created for most other kinds of services.

This is made possible through the use of shared interaction protocols, which can be written by any user of the system and then reused by any peer on the network. These protocols describe interactions between two or more peers, and detail the messages that will be sent as part of that interaction and the constraints on those messages, which give information about the semantics of the messages and under what circumstances they can be sent. A peer can play a role in any of these protocols in which it can satisfy all the relevant constraints.

This shift of emphasis to interaction means that semantic agreement can be reached locally. The lack of a global ontology means that peers will be using different terms which may not match, but interpretation of these terms need only be done within an interaction and only for those terms that directly affect the interaction. The problem of semantic heterogeneity is therefore reduced to the specific case of terms that are actually in use and the matching takes place within a particular context - the context of an interaction - which aids the interpretation process.

To learn how to use OpenKnowledge system, click here.