The list of all past and present tasks for this project can be seen here: http://code.google.com/p/inqle/issues/list?can=1&q=label:Type-Task&sort=-status+milestone+priority&colspec=ID%20Type%20Status%20Priority%20Milestone%20Owner%20Summary

Here is the current to-do list: http://code.google.com/p/inqle/issues/list?can=1&q=label%3AType-Task+status%3AAccepted&sort=-status+milestone+priority&colspec=ID+Type+Status+Priority+Milestone+Owner+Summary&cells=tiles

Below are the milestones, past and future.

Releases

Version 0.1

(released in June 2008)

First release, with basic functionality for local data.

• Eclipse RAP-based user interface, in Equinox OSGi container
• Management of Jena-based RDF databases
• Single sampling algorithm: SimpleSparqlSampler, which extracts example set, in which each row has common set of predicates
• RapidMiner-based experiments
• Jenabean-based Object-RDF persistence
• Full support for running the learning cycle
• Extensible, runnable agent plugin type, including agent for running
• support for customization of LearningCycle and/or its components
• viewing history of experiment results
• embedded H2 database, for simplified installation

Version 0.1.4

(October 1, 2008)

Improved usability, without need to enter URIs.

• New data importer, which captures both explicit and implicit data from spreadsheets, and enables the user to search existing stores of subjects and properties.
• The Central INQLE Server provides a registry for sharing the subjects and properties that other inqle servers have used, permitting standardization and cross-linking of data.
• User name + password required to access the server. Setup wizard captures this admin account info

Version 0.1.5

(October 18, 2008)

• Users can add inferred statements to a dataset, by inputting 1 or more reasoner rules.
• On importing data using the File Data Importer, improved the lookup of subject and predicate URIs. Now using UMBEL subjects and associated schemas (FOAF, etc.). Expernal properties are associated to UMBEL subjects.
• Added new extension type: org.inqle.data.datasetFunctions. This extension can be used by plugin contributors to specify set of datasets, that the plugin can use to do its work. E.g. a plugin developer could create a "trivia" dataset function, and write a "trivia" plugin, which asks trivia questions from whichever datasets of the "trivia" function are available.
• Added dataset function dataset function org.inqle.datasetFunctions.schemas, for lookup of subjects and properties, when importing data.

Version 0.1.8

(25 March, 2008)

• New subject-based sampler available. Agnostic of how the data was loaded. Uses full fledged arcs as attributes
• Fixed bugs in multiple wizards
• Added QName abbreviations of URIs & arcs
• Central INQLE server working for lookup of subjects and some predicates

Version 0.2

(9 December, 2008)

• Above features
• new generation of usability enhancements to file data importer wizard
• Sampling algorithm which is suited data imported via the File Data Importer.
• Usability fixes

Version 0.2.1

(19 December, 2008)

• Added ability to plug in a database
• Added cache database, with improved performance in learning

Version 0.2.2

(23 December, 2008)

• Bug fixes

Version 0.2.3

(13 January, 2009)

• Added deleting of past experiments
• New data listing/browsing views, supporting faceted browsing
• Added clearing of external datasets

Version 0.2.4

(22 January, 2009)

• Added actions to the detail view, and made it scrollable
• Moved dataset browsing to a new view

Version 0.2.5

(5 February, 2009)

• Replaced traditional databases (Jena SDB, JDBC) with a new native RDF database called Jena TDB, with improved performance.
• Switch to TDB also fixed a memory leak, due to Jena SDB running on H2 database.
• Enhanced first data sampler, adding ability to configure minimum and maximum number of learnable attributes to select.
• Reorganized data into 2 categories: system (formerly internal) and user. Added browsing of data in system datamodels. Renamed "dataset" to "datamodel".
• Added multiple RapidMiner performance measures to the data that is stored each experiment.

Version 0.2.6

(6 February, 2009)

• Fixed bug which caused INQLE to die after about 5 hours continuous running.
• changed INQLE namespace to http://ns.inqle.org/1/

Version 0.2.7

(10 February, 2009)

• Caching Jena Models, with big performance gains. INQLE is now running "like a top" - 3.7 seconds per cycle (against 550 records, 3 learnable attributes + label), whereas it was taking 11 seconds using previous database (Jena SDB).
• Central INQLE Server is now running more reliably. Faster performance there as well, so lookups of classes and properties will return quicker

Version 0.2.9

(25 February, 2009)

• Exporting of selected RDF subjects as RDF
• Publishing of smaller datamodels as RDF files

Version 0.4

(expected end of June, 2009)

Enhanced connectivity to data.

• Use of SPARQL endpoints as INQLE datasets
• INQLE servers are SPARQL endpoints
• Sampling algorithms which combine local with remote data
• Survey module: user creates/reuses survey questions, adds participant emails, specifies parameters such as frequency and duration of mailings, and starts the survey.

Version 1.0

(expected late 2009)

Lives up the the name, Intelligent Network of Querying and Learning Engines