Glossaries passolo trados translation memory. Translation Memory Technology. What is “translation memory” and how does it work?

13.11.2021

TRANSLATION MEMORY / TRANSLATION ACCUMULATOR / PROFESSIONAL EDUCATION/ TRANSLATION / TRAINING IN TRANSLATION/ TM-PROGRAM / TRANSLATION MEMORY / PROFESSIONAL EDUCATION / TRANSLATION / TEACHING TRANSLATION / TM SOFTWARE

annotation scientific article on computer and information sciences, author of the scientific work - Olga Ivanovna Babina, Pavel Grigorievich Osminin

The problem of teaching students to work with specialized software is considered. The history of the development of translation memory programs is covered, examples of the most famous programs, including Russian developments, and criteria for their differences are given, the general process of working with programs is described, and the most typical areas of application of translation memory technology are indicated. The advantages of translation memory programs are indicated, which consist in increasing the translator’s productivity and simplifying compliance with the unity of terminology in the translation text. The disadvantages of such programs are given; increasing labor efficiency is only possible by translating texts containing a large number of repetitions, learning to work with programs requires, as a rule, time and financial costs. The application in educational process translation memory technologies, and also provides Various types tasks for students aimed at developing the skill of using translation memory. Assignments reflect the real specifics of a translator’s work, which increases the competitiveness of the future graduate in the labor market. A methodology for assessing learning outcomes is presented. The results of the article can be used in drawing up methodological recommendations for the course translation training.

Text of scientific work on the topic “Translation memory in translator training”

UDC 378.16 BBK Sh11

TRANSLATION MEMORY IN TRAINING TRANSLATORS

O.I. Babina, P.G. Osminin South Ural State University(Chelyabinsk)

The problem of teaching students to work with specialized software is considered. The history of the development of translation memory programs is covered, examples of the most famous programs, including Russian developments, and criteria for their differences are given, the general process of working with programs is described, and the most typical areas of application of translation memory technology are indicated. The advantages of translation memory programs are indicated, which include increasing the translator's productivity and simplifying compliance with the unity of terminology in the translation text. The disadvantages of such programs are given - increasing labor efficiency is only possible when translating texts containing a large number of repetitions; training to work with programs usually requires time and financial costs. The use of translation memory technology in the educational process is substantiated, and various types of tasks for students are given, aimed at developing the skill of using translation memory. Assignments reflect the real specifics of a translator’s work, which increases the competitiveness of the future graduate in the labor market. A methodology for assessing learning outcomes is presented. The results of the article can be used in drawing up methodological recommendations for a translation training course.

Key words: translation memory, translation memory, vocational training, translation, translation training, TM program.

Introduction

In the modern world, the need for translation services is increasing. The volume of the translation market is growing, the need for qualified translators is increasing, and the productivity of translators is required to increase. The role has increased significantly information technologies in translation - wide use got it like general programs(text editors) and specialized translation programs (electronic dictionaries, text corpora, translation memory programs, terminology databases). The material for translation itself has undergone changes - increasingly digital material is offered for translation in the form of presentations and multimedia materials. A number of authors note that when training translators, it is necessary to take into account changes in the work of translators, teach modern skills and abilities, and use modern information technologies in training.

Overview of TM programs

Translation Memory (TM) technology is designed for reuse

in translation activities of previously translated texts.

Some of the first attempts to use a computer to assist in manual translation date back to 1966. In Mannheim, the German Armed Forces Translation Agency used a computer to create glossaries for certain types of texts. The work proceeded as follows: the translator read the English text and underlined the words for which he needed German equivalents. The computer operator then fed this information into the computer, which could compile three or four glossaries for the text in ten minutes of work.

In Luxembourg, the terminology office of the European Coal and Steel Community used an automatic dictionary search for terms. According to the bureau's estimates, the translator spent at least a quarter of his time searching for terms. When working with the system, the translator underlined words unknown to him. Then the entire proposal was entered into the computer, automatic search, and the most similar sentences were printed, as well as words with translation in the order of appearance in the text. The more came

requests from the translator, the more sentences were searched, which increased the likelihood of terms appearing in the correct context. Thus, the bureau was able not only to facilitate the work of the translator, but also to compile specialized glossaries in a short time.

In 1979, Peter Arten came to the conclusion that the texts of the European Commission, which had to be translated into several languages, contained a very large number of repetitions and quotations from various commission documents, so that translators were forced to waste time translating already translated texts. Arten suggested that it is necessary to have a large common central memory with a search function, where all the organization’s documents and their translations will be stored. When typing text, it will be possible to specify the required languages for translation and compare the document with documents in the shared memory, from where the closest translations of sentences of the original text in the required languages will be printed.

Martin Kaye, in a 1980 report, described a proposed translator aid device consisting of a word processor and a dictionary. The text editor consists of two windows - the top one, where the original text is contained, and the bottom one, where the translation is entered. The translator has the ability to highlight a word or phrase in the text and search in the dictionary; it was supposed to be possible to edit entries in the dictionary.

The first commercial TM programs appeared in the early 1990s. In 1984, STAR (an acronym for Software Translation Artwork Recording) was founded in Stein am Rhein, Switzerland. The company was developing software, providing translation services. In 1991, the company released the Transit program for operating system DOS, the program included a terminology management module TermStar.

In 1984, TRADOS (an acronym for TRanslation & Documentation Software) was founded in the German city of Stuttgart and in 1992 the company released the Translator's Workbench program for the DOS operating system. The program included a translation memory, a text editor, and a terminology management module MultiTerm.

In 1992, SDL International, a supplier of linguistic services, was founded in the English city of Maidenhead. The company grew rapidly, acquired smaller companies and became one of the market leaders. In 1998, SDL released its own TM program, SDLX. In 2005, SDL acquired TRADOS.

In 1993, ATRIL was founded in Madrid and released Deja Vu, the first TM software for the operating room. Windows systems 3.1. The program was built into Microsoft Word for Windows 2.0. In 1996-1997 Deja Vu has been redesigned into a program with a separate interface.

In our country there are quite a few developments related to TM programs. The MT2007 TM program has been developed since approximately 2007 by a single developer and is available as a separate application. Development of the program was suspended for some time; the program may be available under the name CATnip. ABBYY offers ABBYY Aligner, a tool for creating translation memories from parallel texts. Russian company PROMT offers a combination of machine translation (MT) and translation memory (TM) technologies.

There are many different TM programs available today. The most popular are Across, Deja Vu, MemoQ, OmegaT, SDL Trados Studio, STAR Transit, Wordfast.

Typically, both paid and free training materials are available on the website of the company that develops the program. Free educational materials in Russian are available for Deja Vu and SDL Trados programs from official distributors programs in Russia and the CIS countries - Okay translation agency and T-Service company.

On modern stage A translation memory is a database that stores text segments in the source language and the target language. The original segment and the translation segment associated with it constitute a Translation Unit (TU). The role of a segment is usually text sentences.

TM programs can be distinguished according to several criteria.

1) Based on technical implementation, we can distinguish local TM programs and those available online. Using

online versions (for example, Google Translator Toolkit and Wordfast Anywhere) work takes place in an Internet browser window. In Fig. Figure 1 shows an example of working with Wordfast Anywhere. At the same time, the results of work (translation memories, glossaries) in the Google Translator Toolkit can be made available to other users.

Local programs may be standalone applications or embedded applications.

are saved in a text editor (for example, Microsoft Word). The translation takes place directly in text editor. The advantage of this approach is that the user works with the familiar editor interface and has access to advanced text editing capabilities. The disadvantage is that support for file formats is limited - usually only text editor formats are available.

When working with separate applications the entire translation process takes place in the application environment. In this case, it may be available additional formats. The disadvantage may be the need to get used to the program interface.

2) By additional functionality. Terminology management programs are often built into TM programs - they allow you to provide general access to terminology databases, store and process terminology. This allows us to ensure the unity of terminology in translation, ensure compliance with the language

policies within a particular enterprise or subject area.

To build a translation memory, you must be able to compare segments of the source text and the translation in order to then enter these pairs into the program database. Therefore, TM programs are often equipped with a function for aligning text segments (usually sentences).

In addition, directly during the translation process to resolve ambiguity

view the use of a particular word in context. TM programs, as a rule, provide the opportunity to build a concordance of a word - find a list of all uses of the word in the context, while simultaneously presenting methods for translating the lexical unit in each case. In fact, this function provides access to a multilingual text corpus, and the search for lexical items is carried out through interfaces similar to those used when working with text corpora. Some parallel corpora for individual pairs of languages with interfaces for compiling a concordance are available online (such as Tran88earcs, WeBiText, Tranlogi, SupbroTerri, and others). Since searching by information resource Already quite a powerful source of linguistic information for a translator, parallel corpora of texts with a search tool for them are sometimes separated into a separate type of translation memory and contrasted with translation memory programs themselves, where the search for repeated

» File Edit Display Tools? & | ^ | o = |<^? | □ | ^ ■=£>] O @ , Q. Ei | lSi ♦«*

100% (TM ENsRU) Private TU

Commercial TM programs, as a rule, have built-in systems for extracting terminology and composition.

adding a glossary of terms (SDL MultiTerm Extract, SDL MultiTerm, PROMT TerM, Mono-Conc Pro, Simple Concordance Program, etc.).

The “raw” text, which includes automatically replaced terms and text segments with their translated counterparts, is subjected to post-editing at the final stage.

3. Quality check: includes a formal check of the completed translation for completeness, grammatical correctness, correct translation of relevant terminology, which can be carried out by the translator himself, (possibly) the customer, as well as at an intermediate link providing communication between the translator and the customer. In this case, the process of recording completed edits in the translation memory is of great importance, otherwise errors may be repeated in subsequent translations.

Thus, when working with a TM program, the content (including the use of a number of functions for automating translation activities), formal and economic aspects of translation are distinguished. The identified aspects are an important component of the use of translation memory in the professional activities of translators. In this regard, each of these aspects should be reflected in the practical training of future translators in the use of TM programs in translation activities. At the same time, it seems necessary to “guide” the future translator sequentially through all stages of the real translation process in simulated conditions, which poses the task of developing meaningfully consistent tasks when learning to use TM programs.

Training in the use of TM programs

When teaching students at a university, we consider it advisable to use online versions of TM programs, since this approach has a number of advantages:

Online TMs are free, which simplifies their introduction into the educational process.

Online TM can be used on various devices (desktops, laptops, tablets) and operating systems (Windows, Linux, Mac OS).

Online TM can be used wherever there is Internet access; there is no need to install the program on a certain number of computers in advance. This approach allows us to significantly expand the number of students trained. For example,

classes can be conducted with students of specialized specialties (linguists, translators), with students of non-core specialties (additional education courses).

When working with a TM program, a translator must know how to use and update the program's database. It must be remembered that any tool works better if used correctly.

The types of tasks performed by students in the classroom are varied and reflect the specifics of the actual work of a translator. Let's list some tasks.

Assignment to compile a corpus of texts. Students are given the parameters of the texts that need to be included in the corpus (name of the subject area, genre, style of texts, dates of creation, etc.), and are asked to create an educational corpus from several texts. This task is due to the opinion of a number of authors that TMs can be used in conjunction with other teaching aids (for example, text corpora), which further increases the effectiveness of the lessons. At the same time, some methods of using a text corpus in teaching a foreign language can be reused when training translators.

In the “soft” version, a certain general corpus of texts available online is proposed as a source of texts (for example, the National Corpus of the Russian Language, British National Corpus, Michigan Corpus of Academic Spoken English, etc.). Many online corpora have text meta-markup that classifies texts according to a number of parameters. In this regard, selecting texts from such a corpus is a fairly simple task; only the process of using search tools for a text corpus can cause difficulty. Therefore, the use of search engines for text corpora available online should be an integral part of the training of a professional translator.

The “hard” option is aimed at solving a more complex problem - an independent search for texts according to parameters on the Internet. This, in turn, poses the task of using the capabilities of Internet search engines, and, therefore, information retrieval and search tools are also necessary for a translator in his professional work.

Created during this task

text corpora can be used to perform subsequent tasks when training translators.

Assignment to compile a glossary of the text. Students receive a text from a certain subject area and compile a glossary of terms with translation from it in order to use this glossary to control the uniformity of terminology in the translation they perform. As mentioned earlier, to select terminology, as a rule, tools for automatically extracting lexical units from texts are used, which can be built into commercial TM programs, but are often absent in online systems. In the learning process, we propose to use a tool for extracting noun phrases, which can be used both for teaching foreign languages and as an integral component of translator training. Familiarization with the methods of using a tool for automatically extracting terminology from texts is a necessary component of the process of learning to use translation memory programs in professional activities.

An extension of such a task can be the task of presenting the resulting glossary of terms in a structured manner, in the form of a lexical ontology. Along with translation memories, terminology ontologies are being developed as an important tool for complying with enterprise language policies. In addition, there are studies on the use and comparison of terminological ontologies of different languages to create a resource applicable both in translation memory technology and in automatic translation of texts.

Text alignment task and

translation. Students receive a text and its translation and perform text alignment to obtain a full translation memory. A variation on this assignment is to use the automated, interactive term alignment procedure.

Task for a preliminary analysis of the effectiveness of using TM. The student is given two texts and informed that the first text is used to create a translation memory, the second text will be translated using the translation memory. The student is asked to analyze how effective the use of translation memory is for a given pair of texts. Examples of such texts are presented in Fig. 2. To maintain the continuity of tasks, the text for creating a TM can be a subcorpus of the corpus of texts used to create a glossary of terms.

This task boils down to identifying repeating blocks in both texts and calculating the proportion of such blocks in the text used for translation. To solve this problem, in the extended version of the task it is possible to use auxiliary automation tools (compiling frequency lists, using programs for extracting terminology from texts, tools for comparative analysis of used lexical units, etc.).

A task to translate a text using TM. Students receive the text and translate it independently using TM. The text is given from the subject area previously used to form a glossary and translation memory, which makes it possible to use the results obtained from previous tasks. Initially, it is proposed to translate the text given during the previous task. Further

Text for creating TM Text for translation using TM

An initial boundary value problem for the Boussinesq-Love equation defined on a graph is studied. A reduction to the abstract Cauchy problem for a second-order Sobolev type equation is carried out. A theorem on the phase space of the original equation is obtained. The paper considers the optimal control of solutions to the initial-final problem for a linear equation of Sobolev type with a (b, p)-sectional operator. The initial-final problem for the Boussinesq-Love equation, which models the longitudinal vibrations of a beam, is considered. A reduction to an abstract initial-final problem for a second-order Sobolev type equation is carried out. Theorems on the unique solvability of the original and abstract problems are obtained

Rice. 2. Example of texts for analyzing the effectiveness of using TM

It is possible to independently search for text from the same subject area and try to translate it using the same translation memory.

A task to compare translations made using TM. The task is similar to the previous one. However, students receive multiple texts from different subject areas. The task can be performed in several ways.

Option 1: students are invited to use the previously created translation memory and “try out” translating texts on various topics with its help.

Option 2: students are asked to independently compile text corpora for the subject areas to which the issued texts for translation are devoted, and use them to create a translation memory. And then translate each text using the translation memory relevant to it.

As a result, students are asked to compare the resulting translations of texts to determine the usefulness of translation memory for translating each of them. In the first option, obviously, using translation memory to translate text from the same subject area for which the TM was compiled will be more effective than when translating texts from other subject areas. In the second case, options are possible. Depending on the proximity of the texts used as a corpus to the text to be translated, the effectiveness of the created TM may vary. In each option, students are asked to explain what is causing the difference in the effectiveness of TM use.

Assessment Methodology

learning outcomes

To check the results of training in the use of TM programs, we have developed the following methodology, based on the principle of error accumulation - incorrectly completed tasks are assessed according to a number of criteria. For final control during training, the following criteria are used:

Ability to work with software;

The quality of the translation performed.

Evaluate according to the first criterion

expedient from a theoretical and practical point of view. The theoretical side of the assessment is carried out in the form of a multiple choice mini-test, which includes general questions on TM technology and specific questions

questions on the computer programs being studied. An incorrect answer is worth one point.

The practical side consists of assessing the ability to use the functions of the TM program being studied when independently performing tasks of various types, described earlier in this article. The result is assessed on a binary scale (task completed / not completed). In case of failure to complete the task, 2 points are added to the student’s cumulative grade.

The threshold error rate at which the test task is considered completed is 40% of the maximum possible number of points awarded for errors (which is equivalent to 60% correctness of the results). So, if the test includes 4 test questions and an assessment of 2 practical tasks, the acceptable error rate is 3 points.

To evaluate the second criterion, there are many approaches, such as a subjective approach, comparison with the original according to certain parameters, and others.

We propose to use an approach based on error weighting, which is used in various translation companies, both foreign, for example Lionbridge, and Russian, for example Logrus, Pa1ex. The basis of this approach is to classify errors depending on how much each error interferes with successful communication. Translation errors are assigned “weights” - numerical measures characterizing the degree of influence of the error on the process of perceiving the text of the message. More serious errors that distort the meaning of the message are rated higher. In Fig. 3 shows an error classifier for use in training.

When checking a translation, the number of errors of each type is counted and the translation quality index TQI is calculated using the following formula:

TQI = (E16 + B2-2 + E3 1) / W 1000, where E1 is the number of gross errors, E2 is the number of medium errors, E3 is the number of minor errors, W is the volume of the original text in word usage.

The acceptability of a translation is determined by the threshold value of the TQI coefficient, which can be set depending on the degree of tolerance of various types of errors in the translation. For educational purposes empirically

E1 (gross error) E2 (medium error) E3 (minor error)

Distortion of the meaning of the original; unjustified omission of text for translation Incorrect translation of a term; inconsistency of terminology in translation; inconsistency with the glossary; non-compliance with the norms of the target language Typos (for example, missing spaces)

Rice. 3. Error classifier

Rice. 4. Example of automation when assessing translation quality

In this way, we established an acceptable quality factor for the final control - less than 24 points, which is equivalent to 1 gross error or 3 average errors or 6 minor errors per 250 word usages. This coefficient can also be used for intermediate control. Obviously, the threshold value of the coefficient can be tightened (decreased) in order to strengthen students’ habit of striving for maximum translation adequacy, or weakened (increased) when assessing students’ work at the initial stage of developing translation skills using TM.

The final score is equal to the sum of points for the ability to work with the software and the quality of the translation performed. The lower the number obtained, the more successful the learning results. We recommend compiling test materials for the final control in such a way that the total score is at least 80% based on the assessment of the translation quality criterion. Thus, when assessing the ability to work with software using 4 theoretical questions on the use of TM and 2 practical tasks on the “pass/fail” scale and assessing the quality of one translation text using the TQI coefficient, the permissible error coefficients are distributed in relation to each other 11 % (3 points) and 89% (24 points).

The proposed methodology for assessing learning outcomes has a number of advantages.

Each student's work is assessed individually. The student can see the types of his mistakes and work to correct them. Moreover, translation quality assessment can be used separately in classes on translation, having previously configured the error classifier for the subject area or text types.

It is easy to automate scoring using the specified method. In Fig. 4 shows an example of such automation when assessing translation quality. The teacher fills out a spreadsheet indicating the number of each type of error. The overall score is displayed automatically.

As noted earlier, large translation companies use similar methods for assessing the work of a translator. Thus, when using this technique, the real working conditions of a translator are simulated, which makes it possible to prepare students for the real requirements of the market.

Conclusion

In conclusion, we note that the use of TM programs, like the use of any information technology, may encounter difficulties caused by the problem of rejection of the new. This problem is especially acute in the mentality of modern Russian society. At the same time, overcoming this inhibitory factor is due to the spirit of the times. According to the results of a study by the translation agency EnRus, one of the main requirements for translation is

chiku - availability and ability to use TM programs. In this regard, the integration into the educational process of methods and methods of knowledge management in translation memory programs is currently a matter of competitiveness of the future translator in the labor market. The use of translation memory technology in teaching is therefore an effective means of developing practical skills in students and preparing them for market demands.

Literature

1. Zhang, Y. Computer-aided translation teaching by means of modern information technology / Y. Zhang // 7th International Conference on Computer Science & Education (ICCSE). -Mudanjiang, 2012. - P. 1628-1632.

2. Porshneva, E.R. Training of professional translators in the light of the new state standard / E.R. Porshneva, I.Yu. Zinovieva // Higher education in Russia. - 2011. - No. 3. - P. 63-69.

3. Canovas, M. Open source software in translator training / M. Canovas, R. Samson // Tradumatica: traduccio i tecnologies de la in-formacio i la comunicacio. - 2011. - No. 9. -P. 46-56.

4. Language and Machines - Computers in Translation and Linguistics: A LPA C report / National Academy of Sciences, National Research Council; John R. Pierce. - Washington, DC, 1966. - 138 p.

5. Arthern, P.J. Machine Translation and Computerized Terminology Systems: A Translator’s Viewpoint / P.J. Arthern // Translating and the Computer, Proceedings of a Seminar / ed.

B.M. Snell. - London; Amsterdam, 1978. -P. 77-108.

6. The Proper Place of Men and Machines in Language Translation: Research report CSL-80-11 / Xerox Palo Alto Research Center; Martin Kay. - Palo Alto, CA, 1980. - 21 p.

7.MT2007. - http://mt2007-cat.ru/.

8. ABBYY Aligner. - http://www.abbyy.ru/aligner/.

9. PROMT company website. - http://www. promt.ru/.

10. Russian site Deja Vu X2. - http://www.tmemory.ru/.

11. Website of the T-Service company. - http://www.tra-service.ru/.

12. Google Translator Toolkit. - http://translate.google.com/toolkit.

13. Wordfast Anywhere. - http://www. freetm. com/.

14. Macklovitch, E. Two Types of Translation Memory / Elliott Macklovitch // Proceedings of the Twenty-Second International Conference Translating and the Computer 22 (16-17 November 2000). - London: Aslib, 2000. - 15 p.

15. Ilner, A. O. Training of translators in Russia and abroad: a comparative analysis / A. O. Ilner // Education and science. -2010. - No. 8. - P. 65-70.

16. Li, Q. Application of CAT in Translation Teaching / Qingjun Li, Likun Zhang // 2010 Second International Workshop on Education Technology and Computer Science (ETCS). -Wuhan, 2010. - Vol. 2. - P. 559-562.

17. Seewald-Heg, Yu. Localization of software, information technology requirements for a translator in the era of globalization / Yu. Seewald-Heg // Vestnik Perm. national researched Polytechnic un-ta. Socio-economic sciences. - 2012. -No. 16. - P. 85-96.

18. Frantzi, K. Automatic Recognition of Multi-Word Terms: the C-value/NC-value Method / K. Frantzi, S. Ananiadou, H. Mima // International Journal on Digital Libraries. - 2000. - No. 3. - P. 115-130.

19. Combining Linguistics with Statistics for Multiword Term Extraction: A Fruitful Association? / Gael Dias, Sylvie Guillore, Jean-Claude Bassano, Jose Gabriel Pereira Lopes // Proceedings of the 6th International Conference Computer-Assisted Information Retrieval (Recherche d"Informations Assistee par Ordinateur) (RIA0"2000) / ed. Joseph-Jean Mariani, Donna Harman. - France: College de France, 2000. -P. 1473-1491.

20. Chen, J. A Multi-word Term Extraction System / Jisong Chen, Chung-Hsing Yeh, Rowena Chau // Proceedings of the 9th Pacific Rim International Conference on Artificial Intelligence 2006: Trends in Artificial Intelligence / ed. Q.Yang,

G. Webb. - Guilin, 2006. - Lecture Notes in Computer Science/ - Vol. 4099. - P. 1160-1165.

21. Sheremetyeva, S. On extracting multiword NP terminology for MT / S. Sheremetyeva // EAMT-2009: Proceedings of the 13th Annual Conference of the European Association for Machine Translation / ed. Lluis Marquez and Harold Somers (14-15 May 2009). - Barcelona, Spain: Universitat Politecnica de Catalunya, 2009. - P. 205-212.

22. Ruiz Yepes, G. Parallel corpora in

translator education / G. Ruiz Yepes // Redit. Revista Electronica de Didactica de la Traduccion y la Interpretacion. - 2011. - No. 7. - P. 65-80.

23. Zheng, W. Design of a CAI System for Translation Teaching / Wang Zheng // Proceedings of International Conference CESM 2011, Part II / ed. S. Lin and X. Huang. - Wuhan, 2011. -P. 457-462.

24. Babina, O.I. Using a corpus of texts for teaching a foreign language / O.I. Babina // Modern higher school: Innovative aspect. - Chelyabinsk: ChGI Publishing House, 2008. - No. 1. - P. 21-24.

25. Sheremetyeva, S.O. On the use of text processing programs for teaching foreign languages / S. O. Sheremetyeva // Bulletin of SUSU. Ser. "Linguistics". -2012. - Vol. 15. - No. 25 (284). - pp. 56-59.

26. Khomutova, T.N. Scientific text: linguocognitive approach / T.N. Khomutova, O.I. Babina // Bulletin of SUSU. Ser. "Linguistics". - 2009. - Issue. 9. - No. 25 (158). -

27. Madsen, B.N. Principles of a system for terminological concept modeling / B.N. Madsen,

H.E. Thomsen, C. Vikner // Proceedings of the 4th International Conference on Language Resources and Evaluation (Lisbon, Portugal). -Lisbon: ELRA, 2004. - P. 15-19.

28. Gluckstad, F.K. Terminological Ontology and Cognitive Processes in Translation / F.K. Gluckstad // Proceedings of the 24th Pacific Asia Conference on Language, Information and Computation / ed. by Ryo Otoguro, Kiyoshi Ishi-kawa, Hiroshi Umemoto, Key Yoshimoto, Yasu-nari Harada (Sendai, Japan, November 4-7, 2010). - Sendai: Tohoku University, 2010. -P. 629-636.

29. Dyumin, N.Yu. Automatic word alignment as a way to build bilingual lexicons / N.Yu. Dyumin // Linguistics in the context of culture: materials of the V international. scientific-practical conf. (Chelyabinsk, November 28-30, 2012) /general. ed. E.V. Kharchenko. - Chelyabinsk: Publishing house. SUSU Center, 2012. - pp. 92-94.

30. Petrosyan, A.E. Mental “blindness” (the roots of insensitivity to new ideas) / A.E. Petrosyan // Sociology of science and technology. - 2012. - T. 3, No. 3. - P. 24-42.

31. Shakhova, N. How to find a good translator / N. Shakhova // Russian Translators Forum, St. Petersburg, September 23-25. 2011 - http://www.enrus.ru/ru/publications/TFRussia_ru s.pdf.

32. Tishin, D. Assessing the quality of translations from TQI to competencies / D. Tishin. -http://www.slideshare.net/justm/tqi.

Babina Olga Ivanovna, Candidate of Philological Sciences, Associate Professor, Associate Professor of the Department of Linguistics and Intercultural Communication, South Ural State University, [email protected].

Osminin Pavel Grigorievich, graduate student of the Department of Linguistics and Intercultural Communication, South Ural State University, [email protected].

Bulletin of the South Ural State University Series “Education. Pedagogy” _____________________________________________2013, vol. 5, no. 3, pp. 98-108

TRANSLATION MEMORY IN TRANSLATORS’ EDUCATION

O.l. Babina, P. G. Osminin

unification in the text in target language) are shown. Some disadvantages of these computer programs (an ability to increase efficiency only within restricted domains in texts containing many repetitions, high price of the software, time for learning to use the software) are also mentioned. The use of translation memory software in education is grounded. Different types of tasks for students aimed at mastering their practical skills in using translation memory software are suggested. These tasks reflect a specific character of the translators’ work, which allows the alumni to gain competitive advantage at the labor market. Procedure for evaluating training results is given. The results of the paper can be used for compiling academic guidelines for translators’ education.

Keywords: translation memory, professional education, translation, teaching translation, TM software.

Babina Olga, Candidate of Philology (PhD), Associate Professor, Associate Professor of the Chair of Linguistics and Intercultural Communication, South Ural State University (Chelyabinsk), babinaoi@ susu.ac.ru.

Osminin Pavel, postgraduate student of the Chair of Linguistics and Intercultural Communication, South Ural State University (Chelyabinsk), [email protected].

Translation memory

Translation memory (TM, sometimes called “Translation Memory”) is a database containing a set of previously translated texts. One entry in such a database corresponds to a “translation unit,” which is usually taken to be one sentence (more rarely, part of a complex sentence, or a paragraph). If the next sentence of the source text exactly matches the sentence stored in the database (exact match), it can be automatically substituted into the translation. The new offer may also differ slightly from the one stored in the database (fuzzy match). Such a sentence can also be inserted into the translation, but the translator will have to make the necessary changes.

In addition to speeding up the translation process of repeated fragments and changes made to already translated texts (for example, new versions of software products or changes in legislation), Translation Memory systems also ensure uniformity in the translation of terminology in the same fragments, which is especially important in technical translation. On the other hand, if a translator regularly substitutes exact matches extracted from translation databases into his translation, without monitoring their use in a new context, the quality of the translated text may deteriorate.

Each specific Translation Memory system stores data in its own format (text format in Wordfast, Access database in Deja Vu), but there is an international standard TMX (Translation Memory eXchange format), which is based on XML and which can be generated by almost everyone PP systems. Thanks to this, the results of translators' work can be exchanged between applications, that is, a translator working with OmegaT can use the translation memory created in TRADOS and vice versa.

Most Translation Memory systems at a minimum support the creation and use of user dictionaries, the creation of new databases based on parallel texts (English alignment), as well as semi-automatic extraction of terminology from original and parallel texts.

List of Translation memory software systems

According to recent reviews of system usage translation memory The most popular systems include.

(less often - part of a complex sentence, or a paragraph). If the translation unit of the source text exactly matches the translation unit stored in the database (exact match, English. exact match), it can be automatically inserted into the translation. The new segment may also differ slightly from the one stored in the database (fuzzy matching). fuzzy match). Such a segment can also be inserted into the translation, but the translator will have to make the necessary changes.

In addition to speeding up the translation process of repeated fragments and changes made to already translated texts (for example, new versions of software products or changes in legislation), PP systems also ensure uniformity in the translation of terminology in the same fragments, which is especially important in technical translation. On the other hand, if a translator regularly substitutes exact matches extracted from translation databases into his translation, without monitoring their use in a new context, the quality of the translated text may deteriorate.

In each specific software system, data is stored in its own format (text format in Wordfast, Access database in Deja Vu), but there is an international standard TMX (eng. Translation Memory eXchange format ), which is based on XML and which can be generated by almost all software systems. Thanks to this, the translations made can be used in different applications, that is, a translator working with OmegaT can use a software created in TRADOS and vice versa.

Most software systems at least support the creation and use of user dictionaries, the creation of new databases based on parallel texts. alignment), as well as semi-automatic extraction of terminology from original and parallel texts.

Popular PP software systems

According to reviews of the use of PP systems, the most popular systems include:

The English Wikipedia has a list comparing the capabilities of various systems.

Translation memory standards and formats

TMX (Translation Memory Exchange Format - Translation Memory Exchange) format. This standard enables interchange between different translation memory providers. TMX is a common format among translators and is best suited for importing and exporting translation memories. The latest version of this format - 1.4b allows you to restore source documents and their translation from a TMX file.
TBX (Termbase Exchange format - Terminology base exchange). This LISA (Localization Industry Association) format is currently being revised and reissued according to ISO 30042. This standard allows for the exchange of terminology, including detailed lexical information. The core basis of TBX is defined by the standards: ISO 12620, ISO 12200 and ISO 16642. ISO 12620 provides a registry of well-defined “data categories” with standardized names that function as data element types or predefined values. ISO 12200 (also known as MARTIF) provides the basis for the TBX framework. ISO 16642 (also known as the Terminological Markup Framework) includes a structural metamodel for Terminology Markup Languages in general.
SRX is designed to improve the TMX format and make translation memory exchange between programs more efficient. The ability to specify the segmentation rules that were used in the previous translation increases the efficiency of identifying segments in the current text with the content of the PP.
GMX GILT stands for Globalization, Internationalization, Localization, and Translation. The GILT Metrics standard consists of three parts: GMX-V for volume metrics, GMX-C for complexity metrics, and GMX-Q for quality metrics. The proposed GILT Metrics standard is aimed at quantifying the scope of work and quality requirements when implementing GILT tasks.
OLIF is an XML-compatible open standard that is used to exchange terminology and lexical data. Although it was originally used as a way to exchange lexical data between proprietary machine translation lexicons, the format has gradually evolved into a more general standard for terminology exchange.
XLIFF (XML Localization Interchange File Format - XML format for mutual exchange during localization), created as a single file format for mutual exchange, which is recognized by all localization software. XLIFF is the translation industry's best way to exchange information in XML format.
TransWS (Translation Web Services) defines the required parameters for calling web services when sending and receiving files and messages related to localization projects. It was conceived as a comprehensive system for automating the localization process using services on the Internet.
xml:tm, this approach to translation memory is based on the concept of text memory, which allows you to combine author memory and translation memory. The xml:tm format was provided to Lisa OSCAR by XML-INTL.

Advantages and disadvantages

Advantages

Reducing the time and amount of translator work
Improved translation consistency, especially when a group of translators are working on the same project.
Increasing profits by increasing the productivity of a translator or group of translators
Improving the quality of services by increasing the accuracy and uniformity of translation of terms, especially in specialized texts.

Flaws

Can make the translation more “dry”; the very essence of the text is lost if the translation using a translation memory is performed by a low-skilled translator
There is often no connection between the sentence/text offered by the program and neighboring sentences and the text as a whole.
The original must be in electronic form
One undetected mistake can spread throughout the entire project
It is necessary to train the program itself, and when changing jobs - perhaps more than once (if employers work with different TM programs)
Not suitable for all types of texts
High cost of licensed software

Literature

Grabovsky V. N. Translation Memory Technology // Bridges. Journal of Translators. 2004. No. 2. - P. 57-62.

Notes

Wikimedia Foundation.

2010.

See what “Translation Memory” is in other dictionaries:

Contents 1 In psychology 2 In computer technology ... Wikipedia

Mosaic depicting ... Wikipedia

Text in one language along with its translation into another language. "Parallel text alignment" is the identification of matching sentences in both halves of a parallel text. Large collections of parallel texts are called... ... Wikipedia

Parallel text (bitext) text in one language along with its translation into another language. "Parallel text alignment" is the identification of matching sentences in both halves of a parallel text. Large meetings... ... Wikipedia

Translation memory (TM, sometimes called “Translation Memory”) is a database containing a set of previously translated texts. One entry in such a database corresponds to a segment or “translation unit” (English... ... Wikipedia

The article talks about a new translator tool - Translation Memory (TM) technology, thanks to which the mechanization of translation activities is carried out.

Neural networks in the work of a translator

Machine translation theories, which have been developed in a slow pace since time immemorial, have received a major boost since the 1970s. This was due to significant advances in the field of modeling intellectual activity. In addition to purely scientific interest, this was explained by the growing role of interlingual communications in the modern world.

In addition to electronic dictionaries and conversational dictionaries, by the mid-1990s, “electronic translators” (they began to be called MT technology) became quite widespread. In principle, “electronic translators” are programs that could process an entire text. True, the output was not quite what was needed, and frankly, it was not at all what was needed. In order for the result of the MT technology to become a truly coherent text, a person had to work a lot on it.

The concept of “machine translation” has become well known. Domestic programs Stylus (now called Promt And Socrates). However, interest in such programs gradually declined after a period of initial curiosity, and now it is low, even though they have been significantly improved. Currently, they are used mainly to familiarize themselves with the content of foreign language sites on the Internet, as well as to read and write emails in a foreign language.

By and large, the work of a translator has not fundamentally changed over the centuries. Yes, computers appeared, sort of fast and convenient typewriters. Yes, “electronic translators” have appeared. However, while MT technologies have improved, they have not become a tool with a wide range of applications and truly save time and effort.

What is Translation Memory technology

Technology has become a new tool for translators Translation_Memory(TM). In the West, TM technology and the translator tools created on its basis - Translation Memory Tools (TMT) - are well known and widely used. A feature of this technology is the mechanization of translation activities, and not its automation in the form in which the creators of MT technology saw it. Moreover, this more modest, at first glance, solution brought immeasurably more practical benefits than the global concept of “electronic translators.”

The difference from electronic dictionaries and other translator tools is that a typical TM class program is based on neural networks, which are capable to a certain extent of simulating the functioning of the human brain when processing data. These networks are capable of learning and analyzing complex data sets that are difficult to process using linear algorithms.

These programs use a fuzzy algorithm to search for words not only in their dictionary forms, but also in other forms, for example, in another case. In addition, they can find phrases in a different word order. A self-organizing artificial neural network is capable of finding patterns and determining relationships between them. The program compares the piece you are currently working on with the contents of the database and provides this information to the display. It can determine that a given fragment is similar to a similar fragment existing in the system's memory, for example, 99%, 74%, or even 20%.

Every translator is familiar with the feeling that the phrase he is currently struggling with has already been encountered somewhere, once upon a time, at least in an approximate form. Moreover, it came across either in the same material, or in some other one translated earlier. It’s just some kind of déjà vu... I would like to look at it, this phrase, at least for reference, to sensibly formulate the one you are working on now...

You start flipping back through the original, rummaging through the bookshelf. And it often happened that the searched phrase was actually found. Most often, when it is too late, after the work on the translation has been completed. I think at such a moment I would give anything for a tool that would help me find the necessary text. Even now, when you type a translation text using a computer rather than on a typewriter, it is always a pity to redo something that was composed with such difficulty. A translator's long-standing dream is to come up with some means that would free him from the need to translate the same phrase several times.

This kind of idea underlies the creation of Translation Memory technology, or simply TM, on the basis of which more or less workable programs began to appear in the late 80s - early 90s. In fact, such a program is a control shell that works with one or another plug-in database (TM) and which I would call a translation bank. TM is a computerized version of a large filing cabinet in which every phrase you have ever translated is stored, in pairs - both in the original language and in the target language. Each such pair is called bilingual. The program will instantly remind you how this phrase was translated last time. Access to information is easy and fast.

Overview of the main programs of the Translation Memory class

Currently, there are several software products using TM technology on the world market. They differ from each other, and sometimes significantly. However, they all share some common functionality.

For example, they usually have a text editor with two parallel windows. One window is intended for the original text, the second for translation. When the original text fragment is displayed in the original window (it can be a sentence, a paragraph, a single word), a search for a similar fragment in the database begins. If there is exactly the same fragment in memory, it is displayed in the translation window automatically. If exactly the same fragment was not found, but something similar was found, then the found fragment is displayed in the program window, indicating the percentage of match.

The translator edits the translation text proposed by the program, bringing it into line with the original, sends it to the database memory and proceeds to the next one. If nothing is found at all, then the translator translates it from scratch, typing the text into the translation window. Having finished working on this fragment, the translator sends it to the database memory and proceeds to the next one. Thus, TM replenishment occurs automatically as work progresses. Immediately after the bilingual has entered the TM, it becomes available to the user. If the TM program is located on a network drive and other users work with it, then the new bilingual immediately becomes available to other translators. After the original fragment and the translation fragment are sent to the database, the next fragment is displayed in the original window.

The first program of this type, called Translation Manager, was developed by IBM. Initially, IBM used Translation Manager for its own needs - for localization (adaptation into foreign languages) of its own software. Translation Manager was then offered to the market as a commercial product, and for some time it had a monopoly in this area. This program is quite easy to use, fast and provides good opportunities for individual user settings. There are separate windows for each of the three main functions - Translation, Translation Memory and Dictionary. The Translation Memory window displays matches found in memory, if any, and the Dictionary window displays terminology from dictionaries developed by IBM. This program also has some inconveniences, for example, a complicated procedure for searching individual words in the Translation Memory.

Another program - SDLX - was developed for its own needs by an English company SDL, engaged in software localization. Each program included in the kit is installed separately. SDL Edit is, in fact, the program itself with which translation is carried out. It has three windows showing the source text, translation and Translation Memory. It should be noted that the document to be translated cannot be imported directly into SDL Edit. It must first be converted using SDL Convert, which is a definite inconvenience.

Program Déjà Vu is popular among translators in Europe due to its flexibility and adaptability. In addition to convenience, Déjà Vu differs from other products in having the highest price/quality ratio. Compared to the two previous programs, the product in question has a number of convenient features. Experts note that from the very beginning this program was developed in contact with its future users. It is known that the latest version of Déjà Vu was tested, in particular, by Russian translators. One of them, Andrey Gerasimov, wrote a positive review of this program in the magazine MultiLingual Computing & Technology

The program is well known in our country and popular in Europe. STAR Transit. In its ideology, it is completely different from other products of the class in question. You have already gained a certain understanding of them, and you know that they all have a central large-capacity database, a repository of translations, in which the necessary data is found instantly (if it is there, of course). As Bill Gates would say: "Information at your fingertips." So, all Translation Memory programs have an electronic catalog cabinet, but Transit does not. Everything is different with him. The source data and translation are saved in text format as files, which the translator places in directories at his discretion. The translator can then simply tell the program which directories and even which files he needs. Transit then forms an associative network of links, with which it works. Despite some advantages, Transit is not widely used in Russia. Ultimately, Promt, the official distributor of Transit, switched to distributing Trados' Translator's Workbench.

But Translator’s Workbench, or simply Trados, quickly became widespread in Russia, first among translation firms, and then among individual translators. This can probably be explained by the fact that it makes it possible to work without leaving Word, which is psychologically more comfortable. Who likes to look at text dotted with all sorts of codes in square brackets? Translator’s Workbench (Trados), compared to other programs, is more convenient for an ordinary translator who is not a programming specialist; it is more understandable than other programs of this type. In particular, the control buttons are equipped with tooltips. Being in a familiar and understandable Word, you can use all its tools (for example, autotext).

This product has a full range of useful tools that are present in other programs only to one degree or another. One of these tools is Analise, which allows you to analyze the original text in advance and find out whether there are (and in what quantity) complete and incomplete matches. It is worth mentioning such a convenient tool as Placeables, which allows, for example, to automatically change numbers in the text if everything else in the translated fragment matches what is found in the translation bank (TM). The automatic function of replacing decimal points and/or commas, time formats, abbreviations, etc. in numbers is very convenient. to the appropriate parameters specified by the translator.

The vast majority of Russian translators using TM technology prefer Translator’s Workbench (Trados).

What do Translation Memory programs have in common?

Summarizing the description of the main software products of the Translation Memory class, it should be noted that they, as a rule, have a number of common functionalities.

Mixing - Alignment

This function allows you to create translation banks (TM) from existing files. From two files - one with the original text and the other with the translation text - a set of bilinguals is created. Opposite the sentence of the source text, its translation is built. And so on for all the files that the translator compiled into a single collection. The degree of convenience of the combination function varies from program to program.

Maintenance

As bilinguals accumulate in the database, their number may increase so much that the contents of the database may go beyond all limits. Some inaccuracies and even errors may appear, especially if the program is used by many users on the network. There will be a need to edit or clarify something, or even simply delete it. To be fair, it must be said that the mentioned program maintenance tool does not completely solve this problem - in practice it is still difficult to deal with the clogging of translation banks.

Terminology Dictionary - Terminology Program

All the programs under consideration have their own terminological dictionaries, which are essentially electronic dictionaries known to all. Terminology dictionaries can import files in dictionary or glossary format, and they can also be updated manually.

Text editor - Document Editor

This is where the translation process actually takes place. For this purpose, all programs have the two windows mentioned above - for the source text and for its translation. All text editors have the ability to set, as a percentage, the degree of similarity between the source text currently being translated and the text in the translation library. The higher you set the similarity percentage, such as 90% or 95%, the less likely it is to find matching text there. However, you can set a smaller percentage, then you may accumulate a lot of material that would be useful at least for reference.

Concordance (Linking word usage with context) - Concordance

This is a very useful feature of all the programs under consideration. It often happens that a term (or combination of terms) can have several meanings or shades of meaning. By selecting a term, you can view in the window all the bilinguals in the collection, and in a variety of contexts. This always makes it easier to choose the most accurate translation of a given term or combination.

Natalya Shakhova, director of the EnRus agency, believes that most TM products relieve the user of the need to get acquainted with various word processors. It is enough to study the features of the text editor of the product itself, and you will be able to control all the variety of texts, whether they are presented in QuarkXPress, FrameMaker or Adobe Illustrator. There is, of course, some catch here: in order to master the next word processor or publishing system, you will need the appropriate filter program, which - what? That's right - it sells for some money! In general, the money issue is the most serious when it comes to TMT. They are not cheap (usually several hundred dollars), and, of course, there is no compatibility. Therefore, if you purchased (and mastered!) a product for one project, then for the next you may need to purchase another product and spend time learning it again.

Natalia Shakhova further notes that the West is already actively discussing the consequences of the widespread dissemination of such programs. One of them is that a new translator loses the opportunity to enter the market. Just as in Russia it is already very difficult for a translator to get an order if he does not have a computer and a printer, so abroad there are problems with employment for translators who do not have a TM program. This phenomenon also has a positive side - by hiring the owner of a TM program, the customer receives some guarantee of his professionalism. But even beginners should have the opportunity to start!

Attitudes towards Translation Memory technology in our country are different.

When the translation agency Phonetics began working with Trados Translator’s Workbench in 1999, naturally, during contacts with customers, we talked about the transition to the new technology. Contrary to expectations, the reaction was wary and sometimes downright negative. It turned out that the idea of machine translation was already compromised by the thoughtless use of programs like Stylus and Socrat in the 90s. Moreover, they were often used by translators, say, not very highly qualified, who did not bother editing what the “electronic translators” produced. However, the users of these programs can also be understood - editing these texts required a lot of time and effort.

Even now, you can sometimes find customers from Russian organizations showing a wary attitude towards any innovations in the field of translation software. At the same time, foreign companies, on the contrary, require the use of Translation Memory class programs. They often have established preferences, mainly Trados Translator’s Workbench, as well as Star Transit. They often have ready-made TMs (translation banks) that they offer to use.

As for translators, they have an understandable interest in Translation Memory technology. Translation agencies are even more understandably interested in them. They deal not so much with fiction, journalism, poetry, etc., but with operating instructions, troubleshooting guides, and software localization, i.e. with those materials where there are repeated or similar fragments of text.

In their current form, Translation Memory programs appear to have reached a ceiling. Further work on their refinement follows the path of improving their existing functionality. This leads to the fact that programs are becoming more and more cumbersome and complex, and this despite the fact that there are no fundamentally new ideas there. The same Translator’s Workbench (Trados) in its latest version has acquired such a set of tools that it is very difficult for an ordinary user to understand them.

The operation of this program requires the presence of a separate employee on the staff of the translation agency. In addition to the user manual, the documentation includes a TRADOS Specialist Guide. In addition, the documentation package includes a Project Management Guide, MultiTerm User Guide and WinAlign User Guide. Even a trained translator will require a lot of effort and time to review hundreds of pages of these documents.

A group of researchers at the University of Maryland, Baltimore County (USA), under the leadership of Professor S. Nirenburg, a well-known specialist in the field of artificial intelligence and machine translation, is searching for new solutions. So far, the developments are at the stage of theoretical research, and most likely, bringing this research to the stage of a commercial product will require considerable time.

It is noteworthy that the market for translator tools is mainly occupied by foreign companies. I would like to hope that this will not go on forever. The potential for theoretical developments by Russian specialists in the field of semantic analysis is very large. Our research in this area has reached a very high level. There is every reason to expect that these developments will be materialized in the form of fundamentally new linguistic technologies that will fully unleash the creative potential of the translator.

Journal of a practicing translator"Bridges", one of the few professional periodicals for translators. The journal publishes articles by translators of various specialties, both oral and written. On the pages of the magazine, translators specializing in various fields of knowledge have the opportunity to discuss current issues, debate with each other, and share the secrets of their professional skills.

The Project Translation Memory (TM) is a repository of source strings and their translations into different languages, which can be used to speed up the translation of the same or similar strings in your project or other projects.

Each translation made in a project is automatically added to the project's translation memory. The project owner or managers can also load Translation Memory into the project if necessary.

Downloading or uploading translation memories

Go to the tab TM&MT point project settings.
Click the button download or download.

You can upload and download TM in the following file formats .tmx, .csv, .xlsx.

If you are loading TM in .csv or .xlsx file formats, map the columns to the appropriate languages in the configuration dialog.

TM binding

To bind any specific TMs for your project, follow these steps:

Go to the tab TM&MT point project settings.
Click **Bind TM**.
Select the required TMs in the dialog box that opens.
Click on the button Save.

You can set the priority of the selected TMs in the same dialog.

Sharing TM

With all your TMs shared, you will be able to pre-translate any of your projects using all the shared TMs. In addition, the editor will show suggestions for translation of TMs from all TMs assigned to projects you own.

Application of translation memory through pre-translation

Pre-translation through TM allows for a minimum of 100% and perfect matches.

Glossaries passolo trados translation memory. Translation Memory Technology. What is “translation memory” and how does it work?

annotation scientific article on computer and information sciences, author of the scientific work - Olga Ivanovna Babina, Pavel Grigorievich Osminin

Related topics scientific works on computer and information sciences, the author of the scientific work is Olga Ivanovna Babina, Pavel Grigorievich Osminin

Text of scientific work on the topic “Translation memory in translator training”

Translation memory

List of Translation memory software systems

Popular PP software systems

Translation memory standards and formats

Advantages and disadvantages

Advantages

Flaws

see also

Literature

Links

Notes

2010.

Downloading or uploading translation memories

TM binding

Sharing TM

Application of translation memory through pre-translation