ENIAC, the first stored program digital computer, is built.
1949
Warren Weaver, Director of the Natural Sciences Division at the Rockefeller Center, sends his Memorandum on Translation to 200 interested people. In it, he mapped out a number of scenarios for the brand new field of “machine translation”. One of these scenarios was inspired by the cryptographic methods that had been extensively developed in the previous 10 years, and was based on “frequencies of letters, letter combinations, intervals between letters and letter combinations, letter patterns, etc. which are to some significant degree independent of the language used.”
The cold war sets in between the West and the Soviet block. It involves a race for knowledge on nuclear science, rocketry and computer technology.
1950
Alan Turing publishes Computation and Intelligence which contains the famous ‘Turing test’: “using language as humans do is a sufficient operational test for intelligence”
1951
Yehosha Bar-Hillel becomes first full time MT researcher at MIT, Boston.
1952
The Georgetown MT Research team is set up under Léon Dostert.
First conference on machine translation. Bar-Hillel stresses the need for syntactic parsing. High quality MT output is considered an impossible goal.
The IBM 701, the first production computer for scientific work, is commercialized.
Christopher Strachey uses a Ferranti random generator on the Manchester University Computer in the UK to generate computerized love poems.
Noam Chomsky publishes the paradigm-shifting book Syntactic Structures, showing how sentences can be characterized by formal generative rules. Indirectly this would influence the design of rule-based MT systems, the first generation of translation technology.
The USAF Automatic Language Translator Mark 1 is ready for work
In Japan, MT research starts at the Electrotechnical Laboratory in MITI’s industrial institute on the first English to Japanese system.
ARPA founded in the US to counter satellite technology innovations in the USSR and build up US technology capabilities. It changed in the 1990s into DARPA.
1959
Bar-Hillel report on “the present state of Automatic Translation of Language” at the first 20 MT centers worldwide.
France creates CETA, a centre for MT research at the Institut Blaise Pascal in Paris.
Emile Delavenay publishes An Introduction to Machine Translation in French in Paris, the first book for the general reader on the topic.
1960
Bar-Hillel publishes his “Demonstration of the Nonfeasability of Fully Automatic High-Quality Translation”
A Linguistics Research Center is founded at the University of Texas to develop what was to become the METAL system. This eventually evolved into the MT software now developed by Lucy Software.
Creation of Association for Machine Translation and Computational Linguistics in USA
Peter Toma sets up Computer Concepts after leaving the Georgetown MT project and develops AUTOTRAN, later to evolve into the Systran system used today.
Ted Nelson creates the term ‘hypertext’ (after trying out linktext, jumptext and zapwrite). It heralded the beginning of a more interlinked vision of digital knowledge.
1964
Creation of Automatic Language Processing Advisory Committee (ALPAC) formed by the National Academy of Sciences in the U.S. to study the feasibility of MT
Work starts on DICAUTOM, later to become EURODICAUTOM, a very large computer based lexicon for European Union terminology.
1965
The TAM MT research group is founded at the University of Montreal, Canada, and eventually leads to the successful TAUM weather MT system.
The US sends its first troops to Vietnam at the start of what became the Vietnam War
The term ‘artificial intelligence’ is coined by John McCarthy at a conference held at Dartmouth College in the USA. The capacity to process natural language (understand and produce discourse) is presumed to be a critical part of artificial intelligence, and stimulated a long stream of research in language, translation and speech processing.
Peter Toma founds LATSEC (Language Automated Translation, System and Electronic Communications) in La Jolla to market Systran – first real commercial MT company.
In Hong Kong, development starts on CULT (Chinese University Language Translator) for Chinese into English M. In 1975 it began translating Chinese mathematical texts.
Terry Winograd writes SHRDLU (the name is based on the order of print repositories in a classic printing house) in the Microworld project at MIT. A user can ask questions and give commands to a robot arm in a block world.
ARPA opens ARPANet, the first packet-switched data communications network in USA, and the forerunner of the Internet.
Speech: John Pierce of Bell Labs says that automatic speech recognition will not be a reality for several decades because it requires “artificial intelligence”.
1970
The French Textile Institute starts using TITUS, a translation automation system used to translate abstracts from and into French, English, German and Spanish. A very early example of an efficient solution for very constrained source domain.
Carnegie Mellon University Computer Science Department developed Harpy, the first real-time 1,000 word continuous speech recognition system.
The first email starts working on ARPANet
Xerox creates PARC (Palo Alto Research Center) to develop office of the future. It led to some of the major breakthroughs in office automation and the desktop environment. Natural language processing was a key component in the R&D program
1971
Logos begins working on a rules-based English-Vietnamese engine so that the US authorities could eventually transfer military technology to the South Vietnamese. It used an IBM 360/67 with an output rate of 80,000 to 100,000 words per hour.
Project begins at Brigham Young University to develop an automated translation system for Mormon texts. Read the article on this MT project in The Telegraph of 11/20/1975. This then evolved into an Interactive Translation System (editor plus dictionary and other resources). Part of the original group ended up working on the Weidner commercial system; others formed ALPS (Automated language Processing Systems) and went to Switzerland to develop an alternate system that prefigures Translation memory technology. This group evolved into the major localization supplier ALPNet, finally amalgamated into SDL in 2003.
Speech: DARPA (Defense Advanced Research Projects Agency) establishes the Speech Understanding Research (SUR) program to develop a computer system that could understand continuous speech. One of the largest programs iof its kind in the history of speech before Verbmobil.
Wang launches the first dedicated office word processor
In Japan, Toshiba launches under the table language processing research towards the development of a Japanese word processor, which is finally launched in 1978.
1973
The US pulls out of Vietnam after nine years of military operations. The Logos MT engine is never used as a production system.
1974
Development of new MT systems at the All-Union Centre for Translation of Scientific and Technical Literature and Documentation, Moscow
1975
The US Air Force develops the QUINCE Chinese-English MT system, and continues working on a German to English MT system
Microsoft is founded by Bill Gates and Paul Allen to market their DOS operating system
The European Commission starts development of an independent Systran English-French MT system.
Logos starts developing and English-Persian MT system to help in the sale of military systems to the Shah of Persia. The Shah was ousted in 1979 and the plan abandoned.
1977
Smart AI Inc. created by John Smart to commercialize systems for writing constrained English documentation, partly to simplify the translation/localization process.
METEO starts translating Canadian weather bulletins Eng-Fr. 45,000 words a day with light post-editing.
Yuri Marchuk launches first Soviet MT program: AMPAR (English to Russian)
Weidner Communications Corporation founded to produce computer assisted translation systems for VAX and later PCs. It first demonstrated its Multilingual Wordprocessing System in Utah in 1978. This Weidner TM type engine was later purchased by SDL in 2001.
The first proposals for the Eurotra program: a brand new multiple-language-pair MT system for the languages of the European Community.
1978
The first edition of the ongoing annual Translating and the Computer Conferences, organized by ASLIB in the UK, was held in London. This has acted as a forum for a broad range of information and experience sharing across the translation industry.
Canadian government plans to launch TAUM Aviation for coastal patrol aircraft documentation but it never went into production and the group was disbanded in 1981
Siemens starts funding LOGOS to develop a German-English system for telecoms manuals.
Siemens later funded MT research at University of Texas (original R&D had stopped in 1975). The Weidner MT engine was one source of ideas for this program.
GETA, the French MT research group, launches its ARIANE-78 MT system, which goes through many deployments before being abandoned in the 2000s.
Systran installed at Euro Commission
Xerox starts using Systran to translate technical manuals
New article-title MT system launched in Japan at the Tsukuba Computer Center.
The Chinese Cang Jie input method for computers based on character structure is Invented by Chu Bang Fu. The method is named after Cangjie, the man usually attributed with the invention of the first writing system of China.
MacDonnell Douglas Corp developed guidelines for large-scale document authoring. The 2,000 words based on preferred verbs used in Navy and Air force documents helped create a technical vocabulary that eventually formed the basis of the AECMA (aviation) Simplified English lexicon
Philips Research Laboratory (NL) launches the ambitious Rosetta project, using a logical interlingua
1981
The Weidner translation system is set up in Utah (USA).
1982
The EUROTRA project for a new EU MT system begins, and lasts until 1992. It never produced an operational production system, but managed to stimulate MT research in many countries, from Greece to Sweden, and from Denmark to Portugal.
Japan launches a new 4-year government backed MT research program. This would lead to an explosion of new EN-JP MT projects among major Japanese manufacturers.
Speech: Dragon Systems founded by the Bakers, former IBM researchers. The start of a long history of practical speech and language technology innovations.
IBM introduces 8-bit ASCII allowing European Latin and Greek alphabets with 256 code elements to be displayed and printed.
1984
Trados, the first company to develop and market translation memory technology is founded in Stuttgart. MultiTerm in 1992, and Translator's Workbench in 1994. In 1997, the company received a major boost when Microsoft decided to use Trados for its internal localization needs. It was acquired by SDL in 2005.
Star AG, today a global localization company, was founded in Switzerland. It has developed numerous technologies including TRANSIT translation memory software.
Fujitsu, the Japanese technology giant, launches its Atlas MT JP/EN system.
Bravice International, a Japanese translation company, acquires the rights to the Weidner MT engine. It later sells the technology to Transcend.
IBM starts a R&D program for MT using so-called “slot” grammars.
The world's first instantaneous translation of speech input by a computer was demonstrated by British Telecom's Research Laboratories
The first Machine Translation Summit a biennial event focused on the MT research community and sponsored by the International Association for Machine Translation) was held in Tokyo.
The Unicode project begins, inspired by engineers from Xerox and Apple, to build a Unicode character encoding standard for computers that is universal (addressing the needs of world languages), uniform (fixed-width codes for efficient access), and unique (bit sequence has only one interpretation into character codes). Version 1 of the encoding was published in 1991. Unicode was to enable effective language localization and globalization for all the world’s languages.
At the Conference on Theoretical and Methodological Issues in Machine Translation (TMI) held at Carnegie Melon University, Peter Brown and his colleagues at IBM “shocked” the audience with their report on experiments with statistical machine translation, using the Canadian Hansard corpus. This marked the start of an ongoing and widespread program to explore the potential of SMT as a translation technology.
1989
Intergraph acquires Weidner technology (DP/Translator, renames it as Transcend)
TRADOS (DE), a software localization firm, develops and markets the first translation memory software.
LISA (Localisation Industry Standards Association) founded in Holland as a platform to develop technology standards
IBM runs an R&D project in statistical rather than rules-based machine translation.
Tim Berners Lee publishes his first proposal for what will become the World Wide Web
1990
First commercial MT system for Russian/English/German-Ukrainian, and first university course on MT by Michael Bleckman at Kharkov State University
An international C-STAR consortium is established to research into speech to speech translation (German, Japanese, US) + affiliates
Speech: Dragon Systems releases DragonDictate, the first retailed speech-to-text system for general-purpose dictation on PCs.
The Human Genome Project is launched in the USA, and lasts for 13 years. The aim is to identify the 20,000-25,000 genes in human DNA, determine the sequences of the 3 billion chemical base pairs that make up human DNA, and store this information in databases.
IBM launches its first PC
In 1982, Japan’s MITI began publicizing its Fifth Generation project, aiming to put Japan into the lead in IT and electronics. MT formed part of this, serving the critical need to help Japanese scientists and engineers acquire knowledge from English publications. Large tech companies such as Fujitsu, Hitachi, Toshiba and NEC all began developing MT systems as part of this long term effort. Japan starts its so-called 5th Generation project at the Institute for New Generation Computing, backed by MITI. The aim is to bring artificial intelligence into the limelight.
1991
NEC demonstrates INTERTALKER, an automatic speech to speech system at Telecom ’91, combining speech recognition, PiVOT MT and speech synthesis for English, Japanese, French and Spanish
The first GSM mobile phone network opens in Finland.
1992
SDL was founded in the UK. It is now one of the world’s three or four largest translation and localization companies. It acquired MT technology from Transcend in 2001 and the Trados translation memory assets in 2005.
Launch in Germany of the very large-scale 8-year government financed Verbmobil project to develop an operational speech to speech translation system. Over the years, partner companies developed a variety of language technology spinoffs.
ATRIL (ES) brings out the first version of its Déjà Vu translation memory software.
The Language Data Consortium (LDC) was founded at the University of Pennsylvania to collect and curate all types of languages data with a grant from the Advanced Research Projects Agency (ARPA).
1993
The web become free for anyone wishing to connect. This launches a massive rush for web real estate, the creation of web sites and companies dedicated to providing services, applications and technologies.
Belgian technology firm Lernaut & Hauspie begins to acquire speech, language and translation technologies (including the Siemens Metal MT software) and investments from Microsoft to reach a market capitalization of $10 billion. The firm imploded, declaring bankruptcy in 2001, and its technology assets were spun off to companies such as Berlitz and Nuance. Its version of Metal technology eventually made its way to Lucy Software.
Speech: Dragon Systems brings out the first consumer speech recognition technology package.
ELRA, (European Language Resources Association), and ELDA, (the Evaluations and Language resources Distribution Agency), are created in parallel in February, with funding from the European Commission. ELRA focused on collecting language data, including some terminological and bilingual data. ELDA handles the commercial and business tasks of licensing the data to language engineering organizations.
Systran offers free translation of small texts on the World Wide Web. Not quite the first public MT service, as Systran offered some translation capability on the French Minitel videotex network in the late 1980s.
Transparent Language acquires Transcend MT technology from Intergraph
Lionbridge, the localization company, is founded. It eventually acquired the translation business of Berlitz and inherited its translation technology, known as Barcelona.
1997
The AltaVista Babelfish service launched on the web using Systran, and notches up 500,000 requests a day.
Speech: Dragon Systems maunches Naturally Speaking, the first continuous dictation system.
1998
Lernout & Hauspie launch iTranslator, the first commercial online MT service. $100 subscription + fee per document. After L&H collapsed, Lionbridge took over the service, still called iTranslator. Not to be confused with the app of the same name using Google Translate available for the iPhone since 2010.
Sakhr launches its first Arabic<->“single byte” language commercial MT system. Read the press release.
ALPNET (eventually integrated into Lionbridge) launches one of the first Language Technology Integration services in the localization industry from its Dublin office. Here‘s the press release.
1999
The first version of Wordfast translation memory software hits the freelance translator market.
First tests of VERBMOBIL, the German-funded speech translation system.
IBM launches its WebSphere translation engine for 8 languages
SDL acquires Transcend and Enterprise Translation Server from Transparent Language
The first speech-to-speech translation Phraselator prototype is tested by the US military in Afghanistan. Several versions would emerge in the following years in various theaters of war.
The National Institute of Science and Technology (NIST) launches its first round of MT system benchmarking, using a series of news stories in various languages to be translated into English. The NIST campaign has since turned into an annual event in which different MT systems, research and commercial, compete using the BLEU score on intelligence type documents. NIST’s main focus is on improving methods used in evaluating MT systems and output quality.
2002
SDL launches SDL Knowledge-based Translation System following successful trials with its client CNH.
Leading the team from ISI, Franz-Josef Och’s engine wins the speed MT competition run by DARPA with a SMT engine. Ochs goes on to become head of Translation Development at Google.
2004
The OpenTrad project funded by the Spanish government begins to develop MT engines for Spain’s various languages. Using an exitsing RBMT engine, the consortium of researchers builds Apertium (for the cognate languages) and Eleka (for Basque). Apertium has evolved into an open source platform for developing multiple engines for less used languages around the globe.
TAUS is founded and holds its first forums.
2005
The German Research Center for Artificial Intelligence (DFKI) begins offering OpenLogos, an open-source derivative of the Logos MT System.
Following the 9/11 terrorist attacks on the US, the American government steps up its interest in and funding for MT as a tool for intelligence collection and analysis. The ‘war on terror’ also prompts work on mobile speech technology and related applications.
The birth and fast growth of the world wide web as a universal communications and display platform.
EuroMatrix, a large scale EC funded effort to develop MT engines for all possible European language pairs is launched, and later extended until 2012. The plan is to integrate all MT models, statistical and linguistic, along with open source tools and data, and learn from the results.
Thailand starts speech-to-speech translation project at NECTEC, to develop a prototype EN-THAI system.
2007
MOSES, the open-source statistical MT engine, is launched as a downloadable kit. It begins to be used in a large scale EU project (Euromatrix) to speed up the MT development of new language pairs.
2008
SDL launches SDL Automated Translation Solutions, a comprehensive online workflow integrating MT and TMs.
NEC launches J-Server Pocket, a text/SMS translation service for mobiles in Japan (JP<->EN, KR, CH)
2009
Apptek launches a hybrid MT system (SMT and RBMT)
Systran releases version 7, a hybrid version of its original RBMT. Includes an automated post(editing module.
Toshiba announces a mobile phone with built-in (not server-based) speech-to-speech translation functionality for English, Japanese and Chinese. Read about it here.
2010
Asia Online launches Language Studio, a comprehensive MT and postediting solution for LSPs
Systran launches its enhanced Enterprise 7 MT software.
Language Weaver launches its ‘quality confidence’ module. The company is acquired by SDL.
IBM launches nFluent, a collaborative corporate translation solution including SMT and postediting, for large corporations.
Google announces a project to develop speech-to-speech translation using its existing MT resources.
Social networking online develops into a massive lifestyle phenomenon. The continual rise and expansion of mobile communications
Cloud computing, enabling massive computing and storage power to be shared on any desktop or device.
