Integrating OpenSource Solutions in your Processes Barcelona June

Integrating Open-Source Solutions in your Processes Barcelona – June 2011 Stefan Pries - Andrä AG Yves Savourel - ENLASO

Why FOSS Should Interest You? Many Proven Solutions • Open source products widely used in many domains o The LAMP stack (Linux, Apache, My. SQL and PHP/Perl/Python) o Lucene, Drupal, Zope, SVN, Firefox, Open. Office, etc. Why not use the ones in the translation domain too?

Why FOSS Should Interest You? They are used already • FOSS projects use FOSS translation tools • Some LSPs do use open-source translation tools • Some translators as well (and not just to translate FOSS projects)

Why FOSS Should Interest You? Maturity • Some projects have been established for a long time • User-driven improvements over the years

Why FOSS Should Interest You? Innovation • The Bazaar model often leads to more creative and innovative solutions • Example: Collaborative translation in FOSS long before it was in the "industry"

Why FOSS Should Interest You? Interoperability • Important part of FOSS projects, because they often are meant to be used along with others FOSS projects • • • Wide support for TMX, PO Increasing support for XLIFF, SRX Web services Documented APIs Command-line interfaces

Some Advantages and Drawbacks + You don't have to pay for software licenses - You still have to spend time to install and maintain things - There is no technical support + There is almost alway a users group that can help (for free) and some projects have a commercial side too - What if the main developers of the project leave? + You still have the source and possibly a community - Developers may not implement what you need + Money can be an incentive, or branching a solution

Let's Explore. . . • Only active projects related to translation • Not even all of them, just a selection Categories: • • • Libraries and tools sets Translation management systems Translation editors Machine translation systems Checkers and QA tools

Demonstrations The best way to discover the tools is to use them. Two scenarios: 1. Integration with proprietary system 2. Integration with other open-source applications

Scenario 2 • Archibald the PM uses Rainbow to pre-translate MIF, Word and XML files from a Global. Sight TM and send them for translation to a Transifex server. • Pedro the translator uses his own Translate Toolkit TM server to provide additional leverage, then uses Omega. T, an Okapi plugin, Hunspell, Language. Tool and Apertium to finish the translation of the Transifex files. • Archibald post-processes the translated documents. 9 different open-source tools

And Many More. . . • • • • Anaphraseus (translation editor in Open. Office) Boltran (online translation editor) Bitext 2 Tmx (aligner) Gaupol (sub-title translation) KBabel (PO editor) Lokalize (translation editor) M 4 Loc (Moses integration tool) POEdit (PO editor) Open Language Tools (XLIFF editor) Open. Logos (MT system) Qt Linguist (translation editor) Tiny. TM (TM server) Zanata (translation management system) …Etc.

Summary • Many different tools available for several tasks o Editing o Quality assurance o Project management o Translation assistance (Terminology, TM, MT, etc. ) • Open standards allow complementing existing processes with open-source solutions • Entire enterprise processes based on open-source tools are possible • Not a solution for everyone, but one worth exploring

For More Information • • • Apertium – apertium. org Global. Sight – globalsight. com Hunspell – hunspell. sourceforge. net Language. Tool – languagetool. org Moses – statmt. org/moses Okapi Framework, Rainbow & Check. Mate – okapi. opentag. com Omega. T – omegat. org Open. TM 2 – opentm 2. org Transifex – transifex. org Translate Toolkit, Pootle & Virtaal – translate. sourceforge. net FLOSS Manual on open translation tools: en. flossmanuals. net/open-translation-tools/

Thanks! Any Questions? stefan. pries@andrae-ag. de ysavourel@translate. com Credits for this presentation: • • Indifex provided the Transifex server access Jordi Serratosa (@jordis) did the Catalan translations Universitat d'Alacant provided the Apertium server access Welocalize provided the Global. Sight server access And many thanks to all the developers of the various open-source projects for their hard work.

Appendix

Libraries and Tools Sets Started in… License Platform / Language Remarks Translate Toolkit October 2003 GPL Independent / Initial build around the Gettext facilities and PO files, Python Okapi Framework Opensourced in November 2005 LGPL Working with more formats now. Works best with software strings (vs. documentation) Integration: Python API, commandline toolset, Web service (tmserver) Independent / Components: filters, connectors, steps, etc. Java Rainbow and Tikal tools Integration: Java API, command-line tools, Rainbow functionalities available through Longhorn Web services

Translate Toolkit

Okapi Framework

Translation Management Systems Started in… License Platform / Language Pootle December 2004 GPL Independent / Existed long before "crowd. Python sourcing" was a term. Example of use: the translation system of Launchpad, the backbone of Ubuntu translations. Transifex Summer GPL 2007 as a GSo. C project Global. Sight (Web-based server application) Opensourced in January 2009 (Web-based server application) Remarks Independent / Mainly PO and TS files Python Integration with many SCMs Integration: API, Web services, command-line client. Apache 2. 0 Windows, Linux / Java Complex system, many features Integration with TDA TMs Integration: Client tool, Web service API

Pootle

Transifex

Global. Sight

Translation Editors Started in… License Platform / Language Remarks Omega. T 2002 GPL Independent / Very active community Java Other projects use Omega. T at their core: e. g. Boltran, Benten Integration: Plugin API, commend-line mode. Virtaal 2007 GPL Independent / Uses Translate Toolkit Python Open. TM 2 Opensourced in June 2010 EPL Windows OS/2 type user-interface LISA wanted to make it a way to provide reference implementation for standards

Omega. T

Virtaal

Open. TM 2

Machine Translation Systems Started in… License Platform / Language Remarks Apertium 2005 (Open. Trad project) GPL Linux, Rule-based system Windows / C++ Some of the language pairs: Catalan<->Spanish, Spanish<->English, , etc. Integration: Web services, command-line Moses 2006 LGPL (Euro. Matrix, etc. ) Linux, Statistical-based system Windows / C++ Other projects are related: - Moses For Localization (M 4 Loc) - Moses for Mere Mortals - Do Moses Yourself CE Integration: Web services, command-line.

Apertium

Moses

Checkers and QA Tools Started in… Language. Tool May 2004 License Platform / Language Remarks LGPL Independent / Grammar and QA checker Rule based, extendable Java Has defaults for many languages Integration: Java API, Web services Hunspell July 2005 GPL, LGPL, Independent / Spell-checker Used in many other tools MPL C++ Integration: API POFilter 2004 GPL Independent / Part of the Translate Toolkit Software string verification Python Integration: Command-line Check. Mate 2010 LGPL Independent / Part of the Okapi project Integration: Java API Java

Language. Tool

Hunspell

PO Filter

Check. Mate