Flyer

English

Apertium (http://www.apertium.org) is a free software (GPL) machine translation platform; it was initially designed to translate between the Romance languages of the Iberian peninsula, but is now being used for more distant pairs.

Who is developing it ?

The Apertium engine is being developed in the Transducens research group at the Department de Llenguatges i Sistemes Informàtics within the Universitat d'Alacant and also by the spin-off company Prompsit Language Engineering. Linguistic data are being developed by Transducens, the Seminario de Lingüística Informàtica of the Universidade de Vigo, the Institut Universitari de Lingüística Aplicada at the Universitat Pompeu Fabra in Barcelona, along with a number of companies including Prompsit Language Engineering, Imaxin|software and Eleka Ingenieritza Linguistikoa, and independent free software developers both in Spain and abroad.

Funding

The Spanish Ministry of Industry, Tourism and Commerce funded the development of the engine and three initial language pairs: Spanish—Catalan, Spanish-Galician and Spanish-Portuguese. The project has also received funding from: the Universitat d'Alacant, the Generalitat de Catalunya (Government of Catalonia) to improve the engine for distant pairs and to develop language pairs such as English-Catalan, Occitan-Catalan and Occitan-Spanish, the Romanian Ministry of Foreign Affairs to develop translators between Spanish-Romanian and Catalan-Romanian.

Currently supported languages

There are currently seven supported translation pairs published using the Apertium platform. These are:

• Spanish-Catalan
• Spanish-Portuguese
• Spanish-Galician
• Catalan-French
• Catalan-Romanian
• Spanish-Romanian
• English-Catalan

Other pairs currently under active development, but without a stable release include: French-Spanish, English-Afrikaans, English-Welsh, Catalan-Romanian, Spanish-Basque and English-Polish. Stable pairs (and unstable ones at your own risk) can be tested through our web interface at http://xixona.dlsi.ua.es/apertium/.

How good is it?

The quality of the final translations depends greatly on the amount of time spent in development, and the closeness of the languages. For example Spanish-Catalan has approximately 95% accuracy, but Spanish-Portuguese has around 90%. For less related and unreleased pairs such as English-Afrikaans, the accuracy, excluding unknown words is somewhere around 70%.

Downloading

Current versions of the engine, linguistic data and documentation can be found on our SourceForge project page (http://www.sf.net/projects/apertium/). Further documentation and discussion can be found both on our wiki (http://xixona.dlsi.ua.es/wiki/) and mailing list (apertium-stuff@lists.sf.net).

Development

The project is always looking for developers who are interested in improving the engine and existing data, working on new language pairs (especially those involving less-used or under-resourced languages), creating interfaces, or adapting the software to fit your needs. Existing free (GPL) data and corpora easily reusable to feed Apertium's dictionaries are also welcome.

Applications

• Multilingual management of web content such as media
• Rapid localisation of free software
• Translation of documentation between a more resourced language and a less resourced language

Македонски

Apertium (http://www.apertium.org) слободна платформа за машински превод на јазици; првично е дизајниран да преведува помеѓу Романски јазици од Иберискиот полуостров, но сега се користи за се подалечни јазици.

Кој го развива ?

Apertium погонот е развиван од Transducens истражувачката група од Department de Llenguatges i Sistemes Informàtics во склоп на Universitat d'Alacant и исто така од компанијата Prompsit Language Engineering. Лингвистичките податоци се развиваат од Transducens, the Seminario de Lingüística Informàtica од Universidade de Vigo, на институтот Universitari de Lingüística Aplicada од Universitat Pompeu Fabra во Barcelona, заедно со поголем број на компании вклучувајќи ги и Prompsit Language Engineering, Imaxin|software и Eleka Ingenieritza Linguistikoa, како и независни развивачи на слободен софтвер - како од Шпанија така и од странство.

Финансирање

Шпанското министерство за индустрија, туризам и комерција го финансираше развојот на погонот и три иницијални јазични парови: Шпанско-Каталонски, Шпанско-Галски и Шпанско-Португалски. Проектот исто така, има добиено средства од: Universitat d'Alacant, Generalitat de Catalunya (Владата на Каталонија) за подобрување на погонот за подалечни парови и за развивање на јазични парови како што се Англиско-Каталонски, Окситански-Каталонски и Окситански-Шпански, Романското министерство за надворешни работи за развивање на Шпанско-Романски и Каталонско-Романски јазик.

Подржани јазици во моментов

Во моментов достапни се седум јазични парови, кои можат да бидат преведувани преку Apertium платформата. Тоа се:

• Шпанско-Каталонски
• Шпанско-Португалски
• Шпанско-Галициски
• Каталонско-Француски
• Каталонско-Окситонски
• Шпанско-Романски
• Англиско-Каталонски

Други парови кои во моментов се во развојна фаза се: Француско-Шпански, Англиско-Африкански, Англиско-Велшки, Каталонско-Романски, Шпанско-Баскиски и Англиско-Полски. Стабилните парови (како и оние во развој, под сопствен ризик) може да бидат тестирани преку нашата веб апликација на http://xixona.dlsi.ua.es/apertium/.

Колку е добар?

Квалитетот на крајниот превод зависи во голема мера од времето поминато во развој и близината на јазиците. На пример Шпанско-Каталонскиот е преведуван приближно со 95% точност, но Шпанско-Португалскиот со околу 90%. За помалку поврзани јазици како што е Англиско-Африкански, точноста е околу 70%(исклучувајќи ги непознатите зборови).

Преземање

Актуелните верзии на погонот, лингивистичките податоци и документацијата се достапни преку SourceForge страната на нашиот проект(http://www.sf.net/projects/apertium/). Понатаму, документација и дискусии може да бидат најдени на нашето вики(http://xixona.dlsi.ua.es/wiki/) и преку мејлинг листата (apertium-stuff@lists.sf.net).

Развој

На проектот секогаш му се потребни програмери кои се заинтересирани до го подобрат погонот и постоечките податоци, работење на нови јазични парови (посебно на оние кои не се користат често или нема доволно ресурси за нив), за креирање на интерфејс програми или адаптирање на софтверот на твоите потреби. Постоечки слободни(GPL) податоци и корпус, кој што лесно може да се вметне во речниците на Apertium се исто така добредојдени.

Употреба

• Повеќејазичен менаџмент на веб содржина
• Брза локализација на слободен софтвер
• Превод на документација помеѓу повеќе застапени и помалку застапени јазици

Castellano

Apertium (http://www.apertium.org) es una plataforma de traducción automática de código abierto (GPL) inicialmente diseñada para las lenguas romances de la Península Ibérica, pero que ha sido recientemente ampliada para poder tratar pares de lenguas más divergentes.

¿Quién lo desarrolla?

El motor de Apertium se desarrolla tanto dentro del grupo de investigación Transducens del Departament de Llenguatges i Sistemes Informàtics de la Universitat d'Alacant como de la spin-off Prompsit Language Engineering. Transducens y Prompsit se encargan también del desarrollo lingüístico junto con el Seminario de Lingüística Informática de la Universidade de Vigo, el Institut Universitari de Lingüística Aplicada de la Universitat Pompeu Fabra de Barcelona y otras empresas como imaxin|software y Eleka Ingeniaritza Linguistikoa. También recibe las colaboraciones de desarrolladores externos voluntarios tanto de dentro como de fuera de España.

Financiación

El Ministerio de Industria, Turismo y Comercio finació parcialmente el desarrollo del motor y de dos de los pares de lenguas iniciales: español-catalán y español-gallego. El proyecto también ha sido financiado por: la Universidad de Alicante (par español-portugués y otros), la Generalitat de Catalunya (mejora del motor para el tratamiento de lenguas distantes y pares inglés-catalán, occitano-catalán, francés catalán y occitano-español), el Ministerio de Asuntos Exteriores de Rumanía (pares español-rumano y catalán-rumano), etc.

Pares de lenguas disponibles

Actualmente hay siete pares de lenguas disponibles que usan la plataforma Apertium:

• Español-Catalán
• Español-Portugués
• Español-Gallego
• Catalán-Francés
• Catalán-Occitano
• Español-Rumano
• Inglés-Catalán

Otros pares de lenguas que están siendo activamente desarrollados pero no poseen aún una versión estable son: francés-español, inglés-afrikáans, inglés-galés, catalán-rumano, español-euskera e inglés-polaco. Los pares estables (e inestables aunque sin garantías) se pueden probar a través de nuestra web en http://xixona.dlsi.ua.es/apertium/.

¿Qué calidad ofrecen?

La calidad de las traducciones finales depende, en gran medida, del tiempo invertido en el desarrollo de un par determinado y la cercanía de las lenguas. Por ejemplo, entre español y catalán se consigue un porcentaje de éxito del 95%; entre español y portugués del 90%. Para lenguas más alejadas y sin versión estable como inglés-afrikáans este porcentaje, sin contar las palabras desconocidas, está alrededor del 70%.

Descargas

Las versiones más recientes del motor, datos lingüísticos, documentación y otras herramientas se pueden descargar de la página del proyecto en SourceForge (http://www.sf.net/projects/apertium/). Se puede encontrar documentación e información adicional tanto en nuestro wiki (http://xixona.dlsi.ua.es/wiki/) como en nuestra lista de distribución (apertium-stuff@lists.sf.net).

Desarrollo

El proyecto busca continuamente desarrolladores intesesados en mejorar el motor y los datos existentes, en trabajar en nuevos pares de lenguas (especialmente aquellos que incluyen lenguas minoritarias o con pocos recursos), en crear interfaces or adaptar el software a necesidades particulares. También se agradece la disponibilización de datos y corpora libres (GPL) que sean reutilizables para mejorar los diccionarios de Apertium.

Aplicaciones

• Gestión de webs con contenidos multilingües usadas, por ejemplo, por medios de comunicación
• Localización rápida de software libre
• Traducción de documentación entre lenguas con muchos recursos y lenguas con pocos recursos

Português

Apertium (http://www.apertium.org) é uma plataforma de tradução automática de código aberto (GPL) que foi projetada inicialmente para traduzir entre línguas românicas da península Ibérica, no entanto atualmente seu uso se expandiu para pares de línguas mais distantes

Quem o desenvolve?

A máquina de Apertium está sendo desenvolvida pelo grupo Transducens formado por pesquisadores do departamento de linguagens e sistemas informáticos da Universidade de Alicante em uma associação com a empresa Prompsit Language Engineering uma spin-off desta mesma universidade.

Linguistic data are being developed by Transducens, the Seminario

de Lingüística Informàtica of the Universidade de Vigo, the Institut Universitari de Lingüística Aplicada at the Universitat Pompeu Fabra in Barcelona, along with a number of companies including Prompsit Language Engineering, Imaxin|software and Eleka Ingenieritza Linguistikoa, and independent free software developers both in Spain and abroad.

Funding

The Spanish Ministry of Industry, Tourism and Commerce funded the development of the engine and three initial language pairs: Spanish—Catalan, Spanish-Galician and Spanish-Portuguese. The project has also received funding from: the Universitat d'Alacant, the Generalitat de Catalunya (Government of Catalonia) to improve the engine for distant pairs and to develop language pairs such as English-Catalan, Occitan-Catalan and Occitan-Spanish, the Romanian Ministry of Foreign Affairs to develop translators between Spanish-Romanian and Catalan-Romanian.

Currently supported languages

There are currently seven supported translation pairs published using the Apertium platform. These are:

• Spanish-Catalan
• Spanish-Portuguese
• Spanish-Galician
• Catalan-French
• Catalan-Romanian
• Spanish-Romanian
• English-Catalan

Other pairs currently under active development, but without a stable release include: French-Spanish, English-Afrikaans, English-Welsh, Catalan-Romanian, Spanish-Basque and English-Polish. Stable pairs (and unstable ones at your own risk) can be tested through our web interface at http://xixona.dlsi.ua.es/apertium/.

How good is it?

The quality of the final translations depends greatly on the amount of time spent in development, and the closeness of the languages. For example Spanish-Catalan has approximately 95% accuracy, but Spanish-Portuguese has around 90%. For less related and unreleased pairs such as English-Afrikaans, the accuracy, excluding unknown words is somewhere around 70%.

Downloading

Current versions of the engine, linguistic data and documentation can be found on our SourceForge project page (http://www.sf.net/projects/apertium/). Further documentation and discussion can be found both on our wiki (http://xixona.dlsi.ua.es/wiki/) and mailing list (apertium-stuff@lists.sf.net).

Development

The project is always looking for developers who are interested in improving the engine and existing data, working on new language pairs (especially those involving less-used or under-resourced languages), creating interfaces, or adapting the software to fit your needs. Existing free (GPL) data and corpora easily reusable to feed Apertium's dictionaries are also welcome.

Applications

• Multilingual management of web content such as media
• Rapid localisation of free software
• Translation of documentation between a more resourced language and a less resourced language

Afrikaans

Apertium (http://www.apertium.org) is vrye sagteware (GPL) vir masjienvertaling. Hoewel dit oorspronklik ontwikkel is om tussen Romaanse tale van die Iberiese Skiereiland te vertaal, word dit tans aangewend vir tale wat verder weg geleë is.

Wie ontwikkel dit?

Die Apertium-enjin word tans ontwikkel deur die Transducens-navorsingsgroep van die Departement van Sagteware en Rekenaarstelsels aan die Universiteit van Alicante, asook by die maatskappy Prompsit Language Engineering wat daaruit ontstaan het. Die linguistiese data word ontwikkel deur Transducens, die Rekenaarlinguistiek-groep (SLI) van die Universiteit van Vigo, die Universitêre Instituut vir Toegepaste Linguistiek van die Pompeu Fabra Universiteit in Barcelona, 'n aantal maatskappye, waaronder Prompsit Language Engineering, Imaxin|software en Eleka Ingenieritza Linguistikoa, en onafhanklike ontwikkelaars van vrye sagteware in Spanje en oorsee.

Befondsing

Die ontwikkeling van die vertaalenjin en die drie aanvanklike taalpare, Spaans-Katalaans, Spaans-Galisies en Spaans-Portugees, is deur die Spaanse Ministerie van Nywerheid, Toerisme en Handel befonds. Die projek het ook fondse ontvang van die Universiteit van Alicante en die regering van Katalonië, om die vertaalenjin vir ander taalpare te verbeter en om taalpare soos Engels-Katalaans, Oksitaans-Katalaans en Oksitaans-Spaans te help ontwikkel, asook van die Roemeense Ministerie van Buitelandse Sake, om vertaalprogramme vir Spaans-Roemeens en Katalaans-Roemeens te help ontwikkel.

Tale waarmee dit tans werk

Die Apertium-platform werk tans met sewe taalpare. Hulle is:

• Spaans-Katalaans
• Spaans-Portugees
• Spaans-Galisies
• Katalaans-Frans
• Katalaans-Oksitaans
• Spaans-Roemeens
• Engels-Katalaans

Ander taalpare wat tans aktief ontwikkel word, maar wat nog nie amptelik beskikbaar gestel is nie, is Frans-Spaans, Engels-Afrikaans, Engels-Wallies, Katalaans-Roemeens, Spaans-Baskies en Engels-Pools. Taalpare met werkende weergawes (asook tale met half-werkende weergawes) kan beproef word by http://xixona.dlsi.ua.es/apertium/.

Hoe goed is die vertaling?

Die gehalte van die eindvertalings hang in 'n groot mate af van die hoeveelheid ontwikkeling wat reeds gedoen is, asook hoe naby die tale aan mekaar verwant is. Die enjin vir Spaans-Katalaans is byvoorbeeld 95% akkuraat, en Spaans-Portugees is omtrent 90% akkuraat. Tale wat nie so ná aan mekaar verwant is nie, byvoorbeeld Engels-Afrikaans, se akkuraatheid is ongeveer 70% (mits alle woorde in die teks bekend is).

Waar om af te laai

Huidige weergawes van die enjin, linguistiese data en dokumentasie kan afgelaai word op ons SourceForge-projekbladsy (http://www.sf.net/projects/apertium/). Ander dokumentasie en vorige gesprekke kan op ons wiki (http://xixona.dlsi.ua.es/wiki/) en poslys (apertium-stuff@lists.sf.net) verkry word.

Ontwikkeling

Die projek verwelkom ontwikkelaars wat graag die enjin en bestaande data wil help verbeter, aan nuwe taalpare wil begin werk (veral tale wat minder algemeen is of waarvoor daar min hulpbronne bestaan), koppelvlakke wil skryf, of die sagteware vir hul eie behoeftes wil aanpas. Bestaande vrye (GPL) data en korpusse wat maklik vir Apertium se woordeboeke aangepas kan word, is ook welkom.

Toepassings

• Veeltalige hantering van webinhoud soos media
• Vinnige lokalisasie van vrye sagteware
• Vertaling van dokumentasie tussen meer gebruikte en minder gebruikte tale