v1.8.1

• Introduced alternative storage backends, mainly to circumvent SDBM's record
  size limit.
  • Warn when truncating a record for SDBM now.
  • Added benchmarks for storage backends.
• Added demo and sample commands to lingoctl and introduced aliases.
• Added default plugin path.
• Fixed Gem dependencies.
• Updated the German system dictionary.
• Extended the README with test setup and literature.
• Internal refactoring and some housekeeping.

v1.8.0

• Ruby 1.8 officially no longer supported! Use Lingo 1.7.x instead.
• Executable may be run from outside installation root directory now.
  See README for further details.
• Search path for dictionaries and configuration files; plugin system.
• Lingo::Attendee::TextReader can read from STDIN if no files given;
  leave out the files parameter in the config for this to trigger.
• Status/performance details as well as progress information are now
  printed on STDERR instead of STDOUT in order to not mix with regular
  output. Learned --log option to write such information to a file.
• Additional internal refactorings. (To be continued...)

v1.7.1

• Lingo::Attendee::MultiWorder actually finds the longest matching multiword
  group.
• Multiword entries in Lingo::Database::Source::SingleWord dictionaries now
  respect def-wc; new parameter def-mul-wc to control
  the word class for multiword entries (defaults to def-wc). Use
  def-mul-wc: 'm' in your .lang to restore the previous
  behaviour.
• New Lingo::Attendee::Formatter for configurable output formatting as an
  alternative to Lingo::Attendee::TextWriter.
• New basic input filters to enable indexing of HTML/XML (and PDF) files.
• Updated the system dictionary.
• Switched license to Affero GPL.

v1.7.0

• More Ruby 1.9 compatibility.
• Extended Lingo::Attendee::VectorFilter to double as enhanced
  Lingo::Attendee::NonewordFilter.
• Fixed Lingo::Attendee::Sequencer to recognize additional word classes beyond
  the first.
• Fixed Lingo::Attendee::Sequencer to not prematurely flush the buffer, which
  led to a decreased recognition rate.
• Housekeeping.

v1.6.12

• Ruby 1.9 compatibility, finally.
• Added lingoweb.rb, a plain & simple web frontend for Lingo.
• Added DIFF-LCS gem directly to Lingo distribution.

v1.6.11

• Redirect Lingo output to /dev/null during testing. Indicate error
  with appropriate exit code.
• Replace usage of command-line diff with cross-plattform Ruby implementation
  (requires diff-lcs).
• Provide alternatives to standard zip command on windows platforms.
• Use UNK itself if it doesn't have any lexicals.
• Use compo form instead of word form when lexicalizing compound entry for
  multiword dictionaries.
• Lingo::Language::LexicalHash#[] must use target (not source) form.
• Optionally, try to find matches for all lexicals a word has.
• Made - a PRINTABLE_CHAR.
• Allow synonyms to be considered for multiword matching.
• Don't use compound parts.
• Introduced some (more or less arbitrary) line length limit. We can only
  store values of a certain length anyway (with SDBM). Entries exceeding this
  limit will be rejected and logged in the .rev file.

v1.6.10: rakefile version 1.6.10

• Prevent "infinite" backtracking in regular expressions.

v1.6.9

• Updated configs.

v1.6.8

• Partly reverted Ruby 1.9 compatibility where it clashed with the 1.8 series.

v1.6.7

• Initiale Bereitstellung von Lingo in Subversion
• BUGFIX: Token wurden durch den Caching-Mechanismus in ihrer Schreibweise
  verändert (Groß-/Kleinschreibung).
• Englische Wörterbücher, ein Beispiel-Artikel und eine Konfiguration für das
  Englische wurden aufgenommen. Aufruf mit lingo.rb -c lingo-en.cfg -l en <datei>.
• Integration der Testumgebung für Entwickler.
• Die Attendees in den Lingo-Konfigurationsdateien werden automatisch
  verkettet. Dies erspart mühsame Verlinkung von Hand und macht die
  Konfiguration lesbarer.
• Modified parameters for Tokenizer and building of databases to handle UTF-8
  correctly. Printable characters from the ASCII charset and every UTF
  character greater than x80 is now recognized as a non word breaking
  character. so now diacritical characters are handled well.
• Added a command line option -p for lingo.rb for making detailed
  performance mesurements possible for each single attendee.
• Allow Lingo.new to be called from a different directory. (Still need to
  Dir.chdir to #talk because of the dictionaries...).
• Allow Textreader to read from STDIN.
• Allow Lingo to be called from an external library or program; pass a string
  and get an array back.
• Allow Textwriter to write to STDOUT.
• Match suffixes case-insensitively.
• Don't exclude synonyms that are eponymous with the original word.
• Don't stop variating hyphenated words if any component was taken as-is.
• Basic Ruby 1.9 compatibility.
• Words should be counted per record, not per document, in LIR mode.
• Added Porter stemmer.