=== 0.4.4
* Namespace usage cleanup
* When serializing graph to RDF/XML, create namespaces as necessary.
* rename bin/reddy to bin/rdf_context

=== 0.4.3
* Coverage tests

=== 0.4.2
* Released as RdfContext gem

=== 0.2.2
* Added AbstractSqlStore and SQLite3Store
* Store contexts are always graphs

=== 0.2.1
* Added MemoryStore, a context-aware store based on Python RDFLib IOMemory
* Add ConjunctiveGraph, a conjunction of all graphs in a context-aware store
* Triple(nil, nil, nil) is wildcard, plus variations. No support for regular expressions.
* BNodes not graph/store specific, generated using milli-second time and sequence.
* Graph(:store) may be one of :list_store or :memory_store, which will instanciate appropriate store
* Graph#contains? iterates over triples if object is a literal to allow for fuzzy matching.
* Change gem name to "gkellogg-reddy" to allow publishing to Github.
* ConjunctiveGraph tests.

=== 0.2.0
* API and RDoc cleanup
* Removed Graph#each_with_subject, Graph#get_resource, Graph#has_bnode_identifier?
* Added Graph#predicates and Graph#objects
* Changed Graph#triples to take optional subject predicate and object to match statements.
* Renamed Graph#bind o Graph#merge!
* Graph#triples yields triples if a block is provided
* Changed Namespace#short to Namespace#prefix
* Change BNode to require a graph. Most uses go from graph.bnode, rather than BNode.new. This is required because bnodes are only equivalent within the same graph
* Added identifier to Graph, defaults to BNode. This from Python RDFlib based on TriX recommendation, see: http://www.w3.org/2004/03/trix/
* Only add triple to graph if it does not already exist in graph.
* Added Graph#contains? to look for existance of triple.
* Added Graph#eql? (and ===) to do simple check for graph equality. This version ignores BNodes (like Python RDFlib)
* Prepend "named" to named BNodes to ensure no overlap between generated and named BNodes.
* Added Graph#clone and Triple#clone
* Merge Graphs, replacing BNodes with new values from new graph.
* Remove duplicate triples when adding or merging graphs
* Added Abstract Store and List Store.
* Delegate Graph methods to store. Currently only supports :list.
* Reduce dependence on HTML namespace in RDFa parser.
* Add RDFa tests for HTML4 and HTML5

=== 0.1.5
* Update RDFa tests to use RdfContext RDF/XML parser.
* All negative RDF/XML parser tests pass.
* Add RdfContext::Parser, super-class of other parsers and move common code.
* Change N3Parser to use same new/parse logic as other parsers.
* Remove debug array by default, and add in calls to tests.
=== 0.1.4
* Update N3 treetop parser to version 1.4.0 (no digits in parser rule names).
* Fix N3 literal handling (still has bug with escaped double quote).
* Handle N3 <#> prefix.
* Don't output lang for typed literals in n3 format
* Rewrite of RDF/XML parser based on spec; passes existing tests.
* Update rdfcore tests based on rdfa test harness.
* Move Literal string escape/unescape to string_hacks.
* Fix literal n3_encoded parsing.
* All positive RDF/XML parser tests pass.

=== 0.1.3
* Added rdfa-test-suite as sub-module and refactored RDFa tests to use them.
* Fix but in white-space separated attributes
* Downcase XHTML reserved words

=== 0.1.2
* Move most test-case logic into rdfa_helper.
* Separate approved and unreviewed tests.
* Allow for soft failure of unreviewed tests by catching Spec::Expectations::ExpectationNotMetError.
* Integrate rdfa-test-suite and create test cases in the same manner. (Causing SPARQL problems with XMLLiterals)

=== 0.1.1 / 2009-11-01
* Changes from rdfa_parser project.

=== 0.1.0 / 2008-12-02

* First public alpha. RDF/XML parsing works but is not fully spec or test compliant. There is much work to be done.