CAT = Computer Assisted TranslationCAT Tools involve translator/computer interaction; they contain no dictionaries:The translation team builds a translation memory database through the translation process of each project. The resulting memory is therefore specific to a particular context (a product and a company).Create an information pool – for all translation data and terminology CAT tools “synchronize” the target language with the source language, unit by unit (i.e., clause by clause or sentence by sentence, rather than word by word).
Translation in its most basic sense is the conversion of words from on language source to another.However, in most cases, you are not seeking to merely translate words, you are looking to translate a concept or an idea.The goal of translation is to transfer semantically and culturally the text and message from one language into anotherTranslation is beyond simple words and focuses on the best way to deliver the meaning behind the words.The success of a translation is highly dependent on the purpose for the translation and the target audience
Glossaries / Terminology Management Systems: Medium for storing standardized terminology (glossary terms, UI etc.)Translation Memory: A linguistic database that stores translation segments. Also referred to as a “translation database.”
The “data” in this case are translation units (TUs) or “segments” that are entered by the translation team, and then stored in a translation memory file. The source text segment is stored & paired with its target text segment. In future translations, the translation memory provides the translation data for any source segment that matches or partially matches the source segment used in the current source translation project.
The computer translates a document using a pre-programmed dictionary and extensive linguistic information.MT is not based on “human translations”: there is minimal or no human involvement.The computer translates a document using a pre-programmed dictionary and extensive linguistic information.MT is not based on “human translations”: there is minimal or no human involvement.
Localization 101 Part 2
Translation/Localization 101:Getting Started – Part II<br />Presented By Anja Schaefer and Arnold Koh <br />January 2010<br />
Glossaries and Terminology Management Systems<br />“The What”<br />Multi-lingual database of key corporate terms, definitions, context, gender, source, etc.<br />Integrated with Translation Memory<br />Establish linguistic standards and encourages consistency in usage<br />Protects corporate terminology and brand<br />Can be shared corporate wide (outside of the translation process)<br />
Glossaries and Terminology Management Systems<br />“The How”<br />Extract of key terms, nouns and noun-phrases from content sources<br />Translations sent to client for approval<br />Approved translations used for all translation projects by translators, editors, proofers, etc.<br />Typically done prior to translation of core materials<br />
Translation Memory<br />Technology that enables users to store translated data within a database for re-use or sharing<br />System matches existing translated segments against new source files<br />Allows for leveraging of translations<br />
Exact Match…is a segment which is 100% identical to a segment stored within the translation memory<br />Fuzzy match… A segment that partially (50-99%) matches a segment within the translation memory<br />Because of the diminishing return as the match gets fuzzier, anything below 75% is considered “no match” for pricing purposes<br />No Match… A segment in the current source text does not <br /> match a segment existing in the TM database<br />Match Types<br />
Translating with Translation Memory<br />This is a sample file to show how a translation memory system works<br />This is the original file<br />It will be translated using Translation Memory<br />This is a sample file to show how a translation memory system works<br />This is the updated file<br />It is not that different to the other file<br />100% match: identical to previous version<br />Fuzzy match: <br />similar to previous version<br />No match: No <br />match found in TM<br />File A<br />File B<br />
Improve consistency and quality<br />Compare new content against existing translated material.<br />Reduce cost<br />Reuse content to reduce and eliminate the need for formatting.<br />Increases speed, productivity, and efficiency of overall process<br />Benefits of Translation Memory and Glossaries<br />
What is Machine Translation (MT)? <br />MT addresses the productivity improvement need for “New Content”<br />MT is characterized by the absence of human intervention during the translation process<br />MT is not and will never be perfect<br />Combining machine and human translation creates a process that ensures the best quality.<br />
Projects Most Suitable for MT<br />Large volume, “repeatable” <br />type content<br />Technical documentation<br />Internal content, “good enough”<br /> vs. publishable quality<br />“Gist” – acceptable quality<br /> (Google/Bablefish)<br />
While not a formal tool or technology, per se, can still be a very important tool for translation<br />Style guides define specific usage, formats, fonts, related to specific localized products<br />Examples:<br />How translators should write titles, headings, numbers, and metrics<br />The manner & tone of addressing the user<br />Items that need to be left in English (fund names, brand references)<br />Standardized treatment for acronyms<br />A Word on Style Guides<br />