During this period of five months, the project team has extended and documented "the WebKB set of tools" (designated "WebKB" below), prepared for publication, published and presented the environment.
WebKB allows its users to store, organize and retrieve knowledge or
document elements (DEs) in Web-accessible files on their local disks or
in a warehouse. The particularity of WebKB is that it unifies different
kinds of complementary tools. It constitutes:
- a private/shared annotation tool of information on the Word Wide Web,
- a knowledge-based Information Retrieval tool,
- a corporate memory (or warehouse) tool, and
- a lightweight Knowledge Acquisition tool.
The core parts for WebKB were developed to allow its users to store, index, organize and retrieve Conceptual Graphs (CGs) based knowledge or DEs in Web-accessible files.
First, the WebKB server (a CGI server) that can execute some document/CG processing commands. The commands may be embedded inside text/HTML documents, and combined with a Unix shell-like script language. The document processing commands are basic Unix-like text processing commands (e.g. cat, grep, fgrep, diff, head, tail, awk) but working on Web-accessible files. The CGs processing commands allow to build and search a CG base. The CG workbenchs CoGITo and Peirce were exploited for that. The DEs may be indexed by CGs and indexed DEs may be retrieved via searches on CGs.
Second, the WebKB interfaces (form-based HTML&Javascript documents) for helping the users to browse ontologies, build CGs, index DEs by CGs and send document/CG processing commands to the WebKB server. Third, Web-accessible documents storing a top-level ontology and various kinds of examples of the use of the WebKB languages and commands. The development of the warehouse has not begun.
March.
Presentation of WebKB to DSTC.
April.
Installation period at Griffith University and on the Gold Coast. Reorganisation
of the WebKB files, amelioration of the user interface, and automatisation
of the procedures of installation of WebKB on different machines.
May.
Beginning of the WebKB general documentation and exploration of the new
litterature and tools in related domains: knowledge acquisition, knowledge
representation, information retrieval, ontology content, ontology servers,
ontology integration, ontology sharing, meta-languages, new languages and
tools proposed by the W3C. Debugging of some functions of the WebKB server,
and extension to allow the loading of several Web-accessible document in
parallel. Second presentation of WebKB to DSTC.
June.
Documentation of the WebKB interfaces and of its representation and query
languages. Writing of a paper with Peter Eklund for the ICIPS'98 conference.
Exploration of litterature on deductive database systems in order to prepare
the building of the warehouse. The idea of using a relational database
system to implement the warehouse was given up for efficiency and inferencing
capabilities reasons. However, no ready-to-use deductive database able
to exploit a large and interactively updatable ontology has yet been found.
Hosted a presentation forum for Melfyn Lloyd and David Barbagello at the
Gold Coast.
July.
Extension of the procedures of cooperative building of an ontology to the
cooperative building of a whole knowledge base. Extension of the WebKB
server for allowing the use of knowledge representation notations less
expressive but more readable than CGs: frame-oriented CGs, HTML-structure
based CGs and indented text. It was decided that no feature other than
the translation of new notations into regular CGs will be implemented in
the CG workbench based WebKB since they would have to be re-implemented
when the warehouse based WebKB is developed.
August.
Preparation for the ICCS'98 conference (talks and demos). Presentation
of WebKB at the ICIPS'98 conference, Gold Coast. Presentation and demos
of WebKB at the ICCS'98 conference, France. Group officially joined DSTC
and developed DSTCII proposal for which WebKB is a research component.
Networking with other researchers, especially researchers of the DSTC, need to be done to find a deductive database able to support the project requirements. Discussions are underway with Steph Crawley about the MOF-based warehouse. As a test bed and a way to allow the procedures for the cooperative building of a knowledge base to be useful in various other contexts, it would be interesting to implement these procedures in the OKBC protocol which allows programs to exploit different kinds of knowledge-based systems, e.g. Ontolingua, LOOM and THEO. Thus, this implementation would also be an extension of each of these systems for the cooperative building of a knowledge base by several users. At this point, publications also need to be written.