coliblog

Today, version 1.0.0 of LanguageTool has been released. So it took about six years from the first implementation in Python (whose version number I cannot remember) to today's release in Java. We fixed quite some bugs, added rules and also added support for three new languages: Danish, Catalan, and Galician. Actually, despite its version number, it's basically a common release.

The latest version of LanguageTool features the following changes:

• Fix for a NullPointerException that could appear when using paragraph-level rules
• Fixes for some bugs with pairing brackets
• Initial support for Icelandic
• More Dutch, English, Polish, Romanian and Slovenian grammar rules
• A new sentence tokenizer that uses the SRX format for specifying end-of-sentence breaks

LanguageTool 0.9.8 adds new rules for several languages (Italian, Romanian, Slovak, Polish, English) and fixes a crash bug. See the change log for details.

We just released LanguageTool 0.9.7. It includes some OpenOffice.org integration fixes and some updates to the error rules. On-thy-fly checking in OpenOffice.org should now be pretty robust (assuming you're using OpenOffice.org 3.0.1). As usual, you can find the details in the change log.

Die fleißige Community hat OpenThesaurus so erweitert, dass es seit heute mehr als 50.000 Wörter umfasst - verteilt auf 18.384 Bedeutungen. Passend dazu hat die Homepage endlich ein übersichtliches Design bekommen: www.openthesaurus.de

We've just released a new version of LanguageTool, the Open Source language checker. It should now work with OpenOffice.org 3.0 without crashing and it supports on-the-fly checking, as I explained in a previous blog entry. Other changes include: more rules for English and Polish and less (but still too many) false alarms for the German agreement checker.

Bei neun gleichzeitigen Tracks (Schedule siehe hier) kann man naturgemäß nur einen kleinen Teil der Vorträge mitbekommen, aber die von mir besuchten Vorträge waren doch sehr interessant. Es ging um Knowledge Managment mit Wikis, Recommendation Systems und die Vorstellung von Startups (z.B. aromicon, Zootool). Leider leidet die Akustik doch sehr, wenn sieben Vorträge in der gleichen Halle stattfinden, getrennt nur durch Stellwände...

Außerdem habe ich gelernt, dass es auch Brezeln ohne Salz gibt. Wenn man bedenkt, wie viel Zeit in unserer Volkswirtschaft durch das Abknibbeln der Salzkörner von Brezeln verloren geht (und es gibt doch wohl niemanden, der das Salz mitisst?), ist das ein echter Fortschritt und ein wichtiger Beitrag zur Lösung der aktuellen Finanzkrise.

• 08.09.2008 2nd Hadoop Get Together Berlin
• 13.09. - 14.09.2008 berlin.jar... die Java-Konferenz in Berlin
• 20.09.2008 motodrone - Internationale Open Source Meisterschaft für Autonome Flugobjekte, in Finowfurt bei Berlin
• 25.09.2008 SuMa-eV Kongress 2008 - Suchmaschinen: IT-politisch, technisch, juristisch
• 30.09. - 02.10.2008 KONVENS 2008 - Konferenz zur Verarbeitung natürlicher Sprache
• 18.10. - 19.10.2008 BarCamp Berlin 3
• 13.11.2008 World Usability Day 2008
• 27.12. - 30.12.2008 The 25th Chaos Communication Congress (25C3)

OpenOffice.org 3.0 is going to be released next month and it features a grammar checker API. That is, a grammar checker can integrate itself into OpenOffice.org and check text while the user is typing text. The most recent release of LanguageTool, 0.9.3, makes use of this, thanks to the work of Marcin Miłkowski. It still has some bugs, but it is useful already:

In related news, the upcoming version of OpenOffice.org will not use the dictionary wizard (DicOOo) anymore, instead dictionaries for spell checking and thesauri will be installed as extensions. You can download those extensions from the official extension website.

Recently, we've started community.languagetool.org, a new website that lets you browse all LanguageTool rules for all languages. The site also checks some Wikipedia articles, finding real grammar errors but also a lot of false alarms which need to be fixed gradually in LanguageTool. We're planning to add more features so that users can share their rules with other users. This way the website will hopefully some day deserve the "community" in its name. Of course, the site is still "beta" as every community-oriented website on this planet.