The Current Status of Development
There are four primary components.
The nenofhat compiler takes human-readable Neno source code and generates a Fhat OWL API. This compiler is written in Java 1.5 using ANTLR. While the compiler is not sophisticated in terms of debugging, it does compile all language constructs.
The second component is the fhat state ontology for the Fhat RVM. This component defines the Fhat RVM instruction set and machine architecture. This component is complete and can be viewed and downloaded at this URL.
The third component is the fhat processor of the Fhat RVM. This processor is written in Lisp and is engineered specifically for the AllegroGraph 2.0+ triple-store. With future development, we hope to provide support for other triple-store implementations.
The forth and final component, is the r-Fhat implementation. This is currently implemented in Lisp for the AllegroGraph triple-store. However, unlike the Fhat RVM, r-Fhat is not tied to any particular triple-store as it is an in-memory representation of the virtual machine and is less of an engineering feat then the Fhat RVM. It is also a significantly faster implementation of the RDF computing model. Because r-Fhat need not be tied to a particular triple-store, a Java implementation will suffice to make it compatible with all systems (though specific interfaces will be required for different web-server and triple-store implementations).
A release is still intended for the Fall of 2007.
Thursday, August 23, 2007
