【正文】 still, but came to life again in the mid 1970s.From mid 1980s to early 1990s, there appeared two epochmaking translation systems KY1 and MT/EC 863, which indicated the significant progress made in China’s machine translation research.From early 1990s to now, China has witnessed unprecedented achievement in machine translation and launched various translation systems such as Transtar, Yaxin, Tongyi etc.Chapter Two Advantages of machine translation Strengths of machine translationSometimes a translator would be required to finish a work within an unrealistically limited time. The only feasible way in the traditional translation may require a team of translators coordinated by a skillful project manager to work concurrently. The alternative solution is to turn to the machine translation software for help. The speed of machine translation can reach 30,000 or even more characters an hour, five to six times faster than humans. Take translation between English and Chinese for example. A professional translator can translate 200 to 500 characters an hour, while machine translation system can finish the same work within several seconds with an accuracy rate as high as about 80%. Evidently, machine translation accelerates the efficiency of translation significantly, saves a lot of time used for consulting dictionaries, and thus reduces the cost of translation.Some machine translation softwares have open database to which users can edit or add monly used jargons. At first it may take some time to enlarge the vocabulary but saves more time in the long term. What’s more, some softwares can also upgrade themselves automatically and expand the vocabulary timely.Geoffrey SamuelssonBrown, a technical Swedish translator has illustrated six major advantages of machine translation in his book A Practical Guide for Translators:l Repetitive or similar texts need to be translated once.l Once glossaries have been entered in the system, future translation will always be consistent providing the translator selects the option offered by the terminology management system.l Greater speed of draft translation, thereby allowing more time for quality control.l A puter can work on draft translation at any time of the day, thus a 10,000 word translation that may take a human translator a week to produce could be done overnight ready for editing the next morning.l Reduction in production costs, thereby producing greater profitability.l Better quality control since text already entered in the software will not need to be rechecked if it can be identified uniquely.(BrownSamuelsson, 2004:77) Various application of machine translation Judging from the quality of current machine translation, there is still a long way to go before realizing the goal of fullautomation and good translation quality. But these machine translation systems are already able to meet people’s diverse needs of translation and are widely used in different fields. From the type of application, machine translation systems fall into the following 4 types: Information distributionThis kind of system is mainly used by information distributors to translate such materials as news, laws, advertisings, product specifications etc. This type of users needs to have information exactly translated into another language, so they require a very high rate of accuracy. To achieve high accuracy, machine translation generally adopts two ways. One is to use restricted language. In the highly restricted field, machine translation can attain quite satisfactory translation quality. The most typical example is the Canadian TAUMMETEO translation system exclusively used in weather forecast. Some panies also use restricted language in their product specifications, which not only ensures that the information contained is clear and easy to understand but also makes machine translation easier. The second way is machineaided human translation. One example in point is translation workbench, a puter aided translation system that adopts the technology of translation memory. Besides translation memory, this kind of product also has the functions of management of terminology database, processing of corpus, management of translation projects and so on. At present, this product has developed into a large industrial scale and is well received by professional translators. The most typical example of this product is Trados system, which has more than 40,000 users and secures 70% of translation software market. Information assimilationThis kind of system serves those who only need to know the general idea but not the exact meaning of the information. This system, which starts developing after the application of the Internet, has attracted a lot of users for it enables a person to understand the basic meaning of websites in foreign languages. One famous example of this is the website of WorldLingo. Information exchange This kind of system targets those who need translation service in onetoone munication. It includes spoken language translation system and text translation system. The former one, spoken language translation system, has made some progress but because of the restrictions of voice technology it’s only used in restricted fields, such as restaurant reservation, customer services etc. The latter one, text language translation system, is free from the problems of voice recognition and is applied in wider areas, such as tourism translation, translation, online chat translation etc. This system has the following characteristics: the material to be translated is mainly spoken language but not written language。 Second, many words are polysemantic and sometimes a sentence may have at least two meanings in a language but puter is not able to decide the exact meaning in the source text. Third, the ways of how sentences are put together are not the same in different languages. These a