MODELER: An Interpreter for WORDS
MODELER is a knowledge representation tool able to map the meaning of
English sentence forms into a Topic Map, exportable in XTM 1.0 format.
This XML file format describes topic symbols and URIs under formal rules
of associations compliant with our Inglish Ontology. They are the conceptual
basis for MODELER, a Java-based interpreter for the WORDS
command language.
Knowledge representation requires procedural components, which
in MODELER appear as O-O scripts embedded in or indexed by related topics. The
WORDS commands they hold work like macros, but map into Java method calls.
WORDS commands may also post to MODELER through network-interfaces. To aid
future expansion, we broadly adopt J2EE specs for its outer UI. This should
let WORDS scripts be triggered easily from browsers, handhelds, and other
networked clients via HTTP, SMTP, SOAP, or custom protocols.
Release 1.0
Conditional on beta tests, this will be released in early '04 by integrating a core
of Lexikos web pages and scripts with these off-the-shelf subsystems:
Portable, net-based MODELER interfaces for humans and other software can
be readily arranged under these APIs. Now being updated each quarter by
Sun, they make your interactive WORDS agents network-accessible by virtually
any client desired. And the XTM paradigm makes it easy to manage and safely
install such agents, even on user demand from a remote WORDS site.
Our core WORDS iinterpreter is a JAR file, backed up by PSIs and pre-built
XTM files from Lexikos that define our top-level ontology. By systematically
extending it, your ontology authors automatically associate its scripts with
their classes, and/or override them with custom replacements that resemble
our examples. Either way, WORDS scripts make them executable objects, able
to constrain, validate, inherit or compute data values as will as contain them.
MODELER initially supports TM4J, one of several free, standards-based TM engines with
Java API's, XML serialization and merging support, query languages, advanced tools,
and good support. They make it easy to emulate a context using dynamically
associated Topics, and find or build new Topics as needed. With XTM import/export,
they can also merge sets of such Topics in as lexicons, or bundle them up as metadata
for return over HTTP to other WORDS users.
|
