Release Notes for Socialtext 2.6.3 - KinoSearch

Socialtext 2.6 includes a major upgrade to the Socialtext search capability. For information about the previous release, see Release Notes for Socialtext 2.2

The new search features and improvements include:


KinoSearch uses stemming, which indexes and searches on the roots of words. For example, if you search for the word "walk", the search engine will also find "walked" and "walking". The stemming feature is currently configured for English. In a future upgrade, we will add the ability to select other languages.

Symbols and other content types

The new search engine includes support for additional content types:

Tag searches and title searches use all of the above features: numbers, alphanumerics, symbols, wildcards, UTF-8, and stemming.

Faster Indexing

Indexing is faster. When new content is added, the new content is available for searching sooner. If the index needs to be completely rebuilt, the re-indexing is much faster as well.

Tag is a synonym for category

tag: is a synonym for category: in the Search box.

Better searching of files uploaded before Socialtext 2.2

The upgrade to Socialtext 2.2, which was released on the Socialtext hosted service in November, added a feature that improved searching of a variety of document types, including Word, Excel, and PDF. This improved searching applied to new files that were uploaded after the upgrade, but did not yet apply to files that were uploaded before the upgrade.

With the upgrade to Socialtext 2.6, older files have been re-indexed. You will now get better results when searching for content in files that were uploaded prior to the Socialtext 2.2 upgrade. For details on the file search features, see Release Notes for Socialtext 2.2.

About KinoSearch

The upgraded search feature replaces the Plucene-based search engine with KinoSearch.

KinoSearch a successor to Plucene. It is a newer, better, and faster Perl implementation of Lucene, a mature and respected Java-based open source search engine widely used in document applications.