GATE Version 5.0 release (May 2009)

1. Major new features

1.1. JAPE language improvements

GATE 5.0 introduces several new extensions to the JAPE language to make it more flexible:

• "Negative" constraints, to support patterns like "match Person annotations, but only where they do not start with an upper-case initial word".
• A richer set of matching operators for feature values, e.g. {Token.length < 5}.
• "Meta-property" accessors, giving access to things like the document text covered by an annotation, e.g. {Organization@string == "IBM"}.
• Contextual operators, allowing you to search for one annotation contained within (or containing) another.
• Bounded range Kleene operators, e.g. ({Token})[2,5] to match between 2 and 5 consecutive tokens.
• Additional operators can be added via runtime configuration.

Montreal transducer users will be familiar with some of these operators, this release brings support for them into the core of GATE.

1.2. Resource configuration via Java 5 Annotations

This release introduces the ability to configure resource classes via source-level annotations, to complement (but not replace) the creole.xml mechanism.

1.3. Ontology-based gazetteer

A new plugin which matches terms taken from an OWL ontology. In combination with a few other generic GATE resources this can produce ontology-aware annotations over text.

1.4. Tools for inter-annotator agreement and merging

Three new plugins to support manual annotation tasks with several annotators working on the same documents. The plugins can copy annotations from several parallel documents into one master document, compute inter-annotator agreement scores between the different annotators, and merge the annotations into a single "consensus" set.

1.5. Packaging applications for GATE Teamware

A menu item and underlying Ant task to assemble a saved GATE application along with all the resource files it uses into a single self-contained package to run on another machine (for example as a GATE Teamware service).

2. GUI improvements

Many elements of the GATE GUI (now called GATE-Developer) have been overhauled to be more user-friendly, with context-sensitive help, a "search and annotate" function in the document editor, drag and drop support in appropriate places, etc. In addition, there is a new schema-driven tool to streamline manual annotation tasks.

3. Other new features and improvements

Many smaller features have also been added. The full list is in the changelog, and includes the following:

• Real-time corpus controller, which terminates processing of a document if it takes longer than a configurable timeout, useful for large batch processes.
• A new sentence splitter based on regular expressions.
• Re-designed ANNIC GUI.

...and many more.

Full details of this release can be found in the changelog. To download GATE, visit the download page.