Summary of current features (v1.3) and the roadmap for planned releases v1.5 & v2.

Features v1.3

• Allows configuration of different content sources.
• Centralized search capability for multiple content sources
• Indexing and searching in Spanish, English, Portuguese, German, French and Nederlands.
• Language auto-detection at indexing time (configurable).
• Configuration of multiple search hierarchies to make precise disctinctions of content by type, location, etc.
• Support for indexing of files embedded in compressed archives:
• ZIP
• TAR
• GZIP
• TGZ
• Viewer Servlet for files inside zip archives
• Search for multiple file types (like the MIME-types configuration of a Web-server)
• Full and incremental indexing, through a cron-like configuration
• Web based administration point.
• API for searching.
• SPI for adding new content sources.
• Advance textual queries (fuzzy, pattern based , etc.). Pattern based search is much like DOS wildcards: * and ? .
• Concurrent indexing and searching.
• One index per content source. Update, enable/disable content sources independently.
• SOAP server for access from non-Java based apps: .NET, PHP, etc.
• Multiple file types support (MIME-TYPES)
• PDF
• MS Word
• MS Powerpoint
• MS Excel
• HTML
• XML
• OpenOffice
• … (see configuration files for the complete list)
• Drivers: FS, FTP, RDBMS (phpcollab, mantis, tutos), OpenCMS, Tree (generic).
• Out-of-the-box way of indexing standard meta-data, e.g.: "Title".
• Search exceptions on Kneobase Webservice mapped to SOAP fault codes = improved error reports in the client side.
• Out-of-the-box installation with a windows auto-installable package

Features v1.5

• Web crawler integration (Nutch? Spindle?)
• On-demand indexing
• Implement standard XML schema for search results (Google Data and/or Open.search)
• Spell checker (~"Did you mean this..."~~)
• Integration/collaboration with other OSS software such as: Solr, Compass
• Upgrade Lucene to 2.0
• New drivers
• SharePoint Driver
• Commons VFS Driver
• NFS Driver
• Pipermail driver (or use web crawler?)
• Test other popular databases (i.e. MS-SQL)
• Support for file embedded in RAR archives
• RDBMS improvements
• implemente getType() for automatic MIME-TYPE
• extend RDBMS driver to reach external documents referenced by a db field
• bundle JDBC most used drivers (Oracle, DB2, SQL Server, Sybase y MySQL).
• Customer satisfaction monitoring when searching
• Queries
• Queries secuence within search session
• Number of hits
• Date and time
• IP address or unique code
• Searching reports for 100 most popular results
• Searching reports for most popular searches that were not found
• Statistics
• Total up time
• Total docs
• Total searches served (? avg per day)
• Domain-specific translation of Spring configuration files

Features v2

• Support for more file types
• Flash?
• Autocad?
• ...
• Dynamic class-loading of parsers and drivers classes (using JMX).
• Cached documents
• Lightweight XML service from JSP (REST approach)
• Direct feedback from user: let them make annotations and content scoring
• Grouping results functionality similar to Google's "Similar Pages" (mainly for forum threads or content with lots of hits)
• Automatic content ranking based on popularity
• Federated network of Kneobase servers
• Fine-grained security
• Online visual configuration editing for each driver
• Driver configuration wizards (notably RDBMS)
• SPI extensions for supporting wizards and online editing
• Thesauri (needs further research, it would be very interesting if Kneobase could provide an automatic classification for browsing knowledge base contents)
• Implement a JMX interface for administration (needs research)
• Save/restore server configuration (drivers/content sources/parsers/builders?)
• Decouple driver configuration and connection
• Versioning of configuration files
• Webstart updates