The Development of New Technologies - Case Study Example

Table of Contents 1.0.Introduction 2 2.0.Methodology 2 Setting the Experiments 3 Experiment 1: Using MemoQ 4 Preview of the experiment 4 Procedure of the experiment 4 Experiment 2: Using Trado 8 Preview of the experiment 8 Procedure of the experiment 8 3.0.Results and Discussion 10 4.0.Conclusion 12 Case Study Comparison of Performance of two Different CAT Tools in the Same Task--- MemoQ and Trado 1.0. Introduction The development of new technologies in the last few years and the explosion of Internet as informative, promotional and communicative tool have modified ways in which translations are done (Zampieri and Vela 2014). As such, technologies have enhanced development of computer aided translation (CAT) tools that not only facilitate processes of translation but provides platform as a basic tool for intercultural communication model that help intercultural understanding. Traditionally, researches in translation have focused on understanding different aspects of CAT tools with studies making comparative analyses between CAT tools and machine translation (Zampieri and Vela 2014). Central hypotheses in different studies are that CAT tools are efficient for translators with such studies holding that they are more consistent and efficient especially when it comes to software localization. However, it has to be understood that these tools offer different methodological alternatives that need to be evaluated. Additionally, there is continued lack of conclusive results that have focused on comparing performance of two different CAT tools in the same task. The aim of this report is to focus on translation memory to provide critical analysis of MemoQ and Trado as CAT tools. Taking experiments through the whole workflow of MemoQ and Trado the report provides a comparative element that involves the two tools on the same task. 2.0. Methodology In order to make translation of the selected text on the two different tools (MemoQ and Trado), the process considered the following principles (Mesa-Lao 2014): The source of language text level focusing on the level of language The cohesive level concentrating on different presuppositions of the second language (SL) text The referential level implicating the level of objects as well as events The level of naturalness basing mainly on reproduction The figure below represents the process of translating the same text using either of the tools. The research process devised experiment based on figure 1 above using MemoQ and Trado in the language pair of Arabic > English with the aim of comparing (developing a comparative element) the two tools in terms of their functionalities, relative usefulness. Setting the Experiments The two experiments were set based on Escartín and Arcedillo (2015) framework where the CAT was considered for the production of the targeted translation from Arabic > English (as two natural languages) without human assistance. The essential distinction of the process is the assessment of syntactic analysis, dictionaries and synthesis components. Just like Escartín and Arcedillo (2015) explained, the comparative analysis will be based on CAT’s abilities in terms of its syntactic analysis, dictionaries and synthesis components. Experiment 1: Using MemoQ Preview of the experiment The first experiment used MemoQ as CAT tool by use the text attached in the appendix. The approach considered MemoQ as multilingual text archive that contain multilingual texts (parsed, segmented, aligned and classified). The tool is taken to have the ability to store and retrieve aligned multilingual segments of the same text against different search conditions that was made. Procedure of the experiment The first step: Creation or opening a project as shown below Second step: Checking the memory of translation (the experiment ensured that there was translation memory where the process would export the translation or alignment results) Third step: Creating a LiveDocs corpus and adding the pair of document Fourth step: The text identified was put into MemoQ (latest version v. 5.0.60/translator pro) and the targeted output generated and stored for comparison. The figure below details this process however, the process separated editable text from the tags by putting aside or splitting the source segments (hotkeys: Ctrl+T). Fifth step: Editing of the alignment results. Sixth step: Exporting of the results to the targeted or primary translation memory Experiment 2: Using Trado Preview of the experiment The second experiment used Trado as CAT tool by using the source text as attached in the appendix. Procedure of the experiment The pre-segmented Trados file was put in MSWord format The source and target of the translation unit was separated by the match statistics which Trados generated (however, the RTF contained the language definitions for the file). The figure below indicates Trados RTF file. The pre-segmented file is imported to Trados as shown in the figure below: Since the text is in Arabic and targeted language is English, the process changed language sections then pressed okay to proceed allowing the text to open with any 100% marches. 3.0. Results and Discussion From the two translated texts (see appendix 2 and 3 for Trado and MemoQ respectively) we have noted that when the same text is translated from Trado and MemoQ, Trado offers platform for auto-completion feature which in turn, provides the process of translation with unique dictionary of aligned phrase that can be extracted from an existing as well as adequately large translation memory (TM). The phrases on the other hand, are suggested to the users as they are typing. The phrase suggestions are what Arenas (2014) terms as ‘prefix’ (p. 47). This advantage indicates that in the past few years, TM developers have been focusing on improving Trado by integrating sub-segment matching termed by Arenas (2014) as ‘concordance’ as well as predictive sub-segment matching which is advanced compared with MemoQ. This approach in turn, shows that Trado is advancing TM leverage more than what MemoQ is offering. As the process of translation was taking place, the auto-suggest and concordance offered by Trado was useful for context and terminology checking but preferring to have the option off was better as it distracted. Differently, the alignment feature provided in MemoQ has been integrated in the software’s LiveDocs component, which offers unique element when working with MemoQ. Accordingly, this feature is a separate resource database where one can put into the system both bilingual and monolingual files that can be used for automatic look up and processes of insertion. In order to ensure that the process of translation is consistent, MemoQ, unlike Trado, provides platform where one can maintain TMs periodically. This is not to mention that it platform is further equipped with an interface where one can correct, edit, delete or add segments. Again, while the two tools offer acceptable levels of consistency, MemoQ has increased levels of consistency owing to the fact that it offers the opportunity for users to recycle existing translation. Just like Zaretskaya (2016) noted, a tool with the ability to recycle existing translation improves the quality of translation, we noted that the best out of MemoQ would be possible when its database is maintained. Instances of perfect matching or exact matching were well elaborated more in MemoQ compared to Trado. While translating part or section of the text that was already translated before, MemoQ was able to suggest the old sentence or segment within its database. The difference between the two tools was that when the translation was using MemoQ the exact match could occur when the term from new second language was identical to old term, provided it was found in MemoQ TM database. The two tools differ in terms of their active terminology recognition. Generally, the two tools provide an opportunity where translator can take advantage of different terminology databases which they create. According to Weitz (2017), this means that translators are given the opportunity to connect the TM to the tools’ ‘termbase.’ However, MemoQ provides users with the ability to manage multilingual glossaries. The tool is designed in a way that it can maintain multilingual glossaries in what it is termed as ‘termbases’ the tool has connected TM. While MemoQ provides users with platform for managing multilingual glossaries, Trado on the other hand, provides users with terminology management systems. Looking at the translated texts from the two tools, Trado provides some levels of identification of equivalence linked with specific field. While the text translated was related to the field of technology and software, Trado offers options for other fields such as medicine, computer, and law among others. Weitz (2017) noted that it would be time consuming for any translation tool searching for every specific term. However, with Trado, the advanced feature for terminology management systems ensures storage, consistency, retrieval and updating entries of terms. Other advantages inherent in Trado include the following: Provides platform for identification of different equivalents Maintaining and organising a termbase Manipulation of terminology resources 4.0. Conclusion The study evaluated the experience in translating a text in both MemoQ and Trado. Based on the analysis, the research has noted that both tools offer platform for flexibility, fast database searching, robustness and wide support of file formats. We have noted that MemoQ provides robustness as compared to Trado making it suitable especially when working with text with frequent repetition of words such as technical texts. These findings show that translation memory tools are becoming useful and indispensable tool for processes of translation. This report further supports previous researches that computer assisted tools help translators in decision-making with specific program they should choose. Bibliography Arenas, A.G., 2014. Correlations between productivity and quality when post-editing in a professional context. Machine translation, 28(3-4), pp.165-186. Escartín, C.P. and Arcedillo, M., 2015. A fuzzier approach to machine translation evaluation: A pilot study on post-editing productivity and automated metrics in commercial settings. ACL-IJCNLP 2015, p.40. Mesa-Lao, B., 2014, April. Speech-Enabled Computer-Aided Translation: A Satisfaction Survey with Post-Editor Trainees. In Workshop on Humans and Computer-Assisted Translation (pp. 99-103). Weitz, M., 2017. Improving retrieval performance of translation memories using morphosyntactic analyses and generalized suffix arrays. Machine Translation, pp.1-30. Zampieri, M. and Vela, M., 2014, April. Quantifying the influence of MT output in the translators’ performance: a case study in technical translation. In Proceedings of the EACL Workshop on Humans and Computer-assisted Translation (HaCat) (pp. 93-98). Zaretskaya, A., 2016. A quantitative method for evaluation of CAT tools based on user preferences. Beyond the universe of Languages for Specific Purposes: The 21st century perspective, p.153. Read More

The first experiment used MemoQ as a CAT tool by using the text attached in the appendix. The approach considered MemoQ as a multilingual text archive that contains multilingual texts (parsed, segmented, aligned, and classified). The tool is taken to have the ability to store and retrieve aligned multilingual segments of the same test against different search conditions that were made.

The second experiment used Trado as a CAT tool by using the source text as attached in the appendix. The pre-segmented Trados file was put in MSWord format

The source and target of the translation unit were separated by the match statistics which Trados generated (however, the RTF contained the language definitions for the file). The figure below indicates the Trados RTF file.

From the two translated texts (see appendix 2 and 3 for Trado and MemoQ respectively) we have noted that when the same text is translated from Trados and MemoQ, Trados offers a platform for auto-completion feature which in turn, provides the process of translation with a unique dictionary of aligned phrase that can be extracted from an existing as well as adequately large translation memory (TM). The phrases, on the other hand, are suggested to the users as they are typing. The phrase suggestions are what Arenas (2014) terms as ‘prefix’ (p. 47). This advantage indicates that in the past few years,  TM developers have been focusing on improving Trado by integrating sub-segment matching termed by Arenas (2014) as ‘concordance’ as well as predictive sub-segment matching which is advanced compared with MemoQ. This approach, in turn, shows that Trado is advancing TM leverage more than what MemoQ is offering. As the process of translation was taking place, the auto-suggest and concordance offered by Trado were useful for context and terminology checking but preferring to have the option off was better as it distracted.

Differently, the alignment feature provided in MemoQ has been integrated into the software’s LiveDocs component, which offers a unique element when working with MemoQ. Accordingly, this feature is a separate resource database where one can put into the system both bilingual and monolingual files that can be used for automatic lookup and processes of insertion. To ensure that the process of translation is consistent, MemoQ, unlike Trado, provides a platform where one can maintain TMs periodically. This is not to mention that its platform is further equipped with an interface where one can correct, edit, delete or add segments. Again, while the two tools offer acceptable levels of consistency, MemoQ has increased levels of consistency because it offers the opportunity for users to recycle existing translations. Just like Zaretskaya (2016) noted, a tool with the ability to recycle existing translation improves the quality of translation, we noted that the best out of MemoQ would be possible when its database is maintained. Instances of perfect matching or exact matching were well elaborated more in MemoQ compared to Trado. While translating part or section of the text that was already translated before, MemoQ was able to suggest the old sentence or segment within its database. The difference between the two tools was that when the translation was using MemoQ the exact match could occur when the term from the new second language was identical to the old term, provided it was found in the MemoQ TM database. The two tools differ in terms of their active terminology recognition. Generally, the two tools provide an opportunity where the translator can take advantage of different terminology databases which they create. According to Weitz (2017), this means that translators are allowed to connect the TM to the tools ‘termbase.’ However, MemoQ provides users with the ability to manage multilingual glossaries. The tool is designed in a way that it can maintain multilingual glossaries in what it is termed as ‘termbases’ the tool has connected TM. 

Read More