Funnelback 9.0.0 release notes

Released: 11th January 2010

New Features

  • New analytics (reports) system: improved UI, geolocation, related queries, emailed reports and improved performance.

  • Query spike detection (detect "spikes" in activity for particular queries).

  • Improved spelling suggestions.

  • Support for 64-bit Windows and Linux platforms.

Upgrade Issues

  • As of version 9.0.0 the Funnelback administration system cannot be run through Apache or IIS and must be run within the embedded Jetty web server. Apache and IIS continue to be supported for the public search interface.

  • Funnelback now requires ActivePerl 5.8.8 or 5.8.9 on all platforms (ActivePerl was previously required only on Windows).

  • Funnelback is no longer supported on Solaris based platforms.

  • Funnelback administrator users (other than the one created during installation) previously created under an IIS based administration interface must be manually recreated, as their passwords are now defined by the jetty .htpasswd file.

  • JDBC drivers must now be installed in <install_root>/lib/java/dbgather/, not <install_root>/lib/java/ as was previously supported. See db.jdbc_class for details.

  • spell.cgi is no longer supported. Spelling suggestions are now returned as part of the results from xml.cgi, and these should be used instead.

Selected improvements and bugfixes

  • Click feedback information now incorporated into ranking by default for new collections

  • Improvements to the display of document titles when no specific title was present in the original document

  • Near-duplicate detection to down-weight repeated occurrences of highly similar results

  • Improved customization options for faceted navigation

  • Support for crawling sitemap.xml files

  • Improved performance for contextual navigation (Fluster)

  • Numerous contextual navigation bugfixes and improvements

  • Administration interface automatically refreshes when collection updates complete

  • Query independent evidence support for meta collections and instant updates

  • Improved support for numerical metadata values

  • HostnameFill option for faceted navigation

  • General availability of document at a time (DAAT) processing for large collection handling

  • Redesigned default search forms, with integrated advanced search functionality

  • Fix issue where adding a new file manager rule could delete existing rules.

  • Fix filtering bug that created invalid XML files.

  • Fix NullPointerException in funnelback-dbgather.jar.

  • Fix funnelback-dbgather.jar to work when large amounts of memory are required.

  • funnelback-dbgather.jar now only outputs valid XML.

  • Robustness fixes for form parsing.

  • QIE fixed to work with meta collections and instant updates.

  • CPU usage when filtering is now throttled, to reduce the load on the server.

  • Filecopy collections ignore temporary files.