Release Notes
=============

Jena 2.0 (August 2003)

This is Jena 2.0.  This version follows on from preview and beta releases
and is the first full release of Jena2.

Most functionality from Jena1 (core RDF API, readers and writers for
RDF/XML, N-Triples and N3,together with RDQL) is provided in Jena2 although
no BerkeleyDB persistence is currently provided.  Jena2 adds new
functionality with an ontology API, including OWL, a reasoner plug-in
architecture together with rule-based reasoners for RDFS and OWL.



Changes from Jena2-beta1:

The jar file "antlr.debug.jar" has been replaced by "antlr.jar".


Persistent Storage:

1) Fastpath is enabled.
2) URI prefix compression is implemented.
3) Configuration and access options are exposed to applications through
   get/set methods.
4) The prefix for database table names may be redefined.
5) Oracle is now fully supported.
6) New factory methods for persistent models are available, deprecating
   the existing ModelRDB constructors.

Reasoning subsystem:
1) Replaced the backward chaining reasoner by a more stable, more space
   efficient LP engine (loosely based on SLG-WAM).
2) Updated the RDFS rule sets as a result, with x2 speed improvement on
   test data.
3) Updated the OWL rule set to be more complete, fix all reported bugs, add
   support for hasValue. More development still required.
4) Added a mechanism for rule-based validation and added validation rules to
   the OWL rule set.
5) Replaced the documentation completely.
6) Fixed various misc bugs report on jena-dev and CVS - many thanks to all
   those took to time to report these problems.

RDF API:
1) Replaced the createTypedLiteral calls with versions that do not allow
   xml:lang tags, in keeping with working group changes.
   Deprecated the version of createLiteral that takes both xml:lang and
xml:wellFormed.



PrefixMappings. Jena models maintain a mapping from namespace
prefixes to namespaces. This mapping is updated when models are
read, and used when models are written. Jena code can query and
update these mappings freely.

Event listeners. Listeners can be attached to models and are notified
when triples are added to or removed from the model, singly or in
bulk.

Reification. Jena optimises the storage of complete reified statements, and
provides classes and methods for manipulating ReifiedStatement objects. By
default, the treatment of reified statments is in accordance with the
latest Working Group specifications.

Model construction. Models can be constructed in memory or in
relational databases. Memory-based models may be reflected in
files. Models can be named and held in clusters using ModelMakers.