Scan: The cornerstone of translation re-use
During the course of a translation, translators often encounter sentences, phrases, or single words that they know they've translated previously. But retrieving the exact corresponding translations often involves searching through old files or even printed documents. Using Déjà Vu X's scan feature, they can instantly check not only how a specific sentence or phrase was translated in the past, but whether a similar sentence or phrase was also translated. Déjà Vu X will display all exact or fuzzy matches for the current sentence, highlighting the differences between the sentence being sought and the proposed matches.
Besides displaying exact and fuzzy matches, Déjà Vu X has a unique feature: self-repairing fuzzy matches or "example-based machine translation (EBMT)." In many cases, when Déjà Vu X finds a fuzzy match for a sentence, it will intelligently evaluate the differences between the two sentences and modify the translation accordingly, turning it into an exact match.
Implementing an effective TM system requires achieving a delicate balance of precision and recall. Using another form of information retrieval—an Internet search engine—as an example, we can define precision as the ability to retrieve only those documents that are strictly relevant to the query; recall is the ability to retrieve a large quantity of documents (which may not all necessarily be relevant to the query). Most systems are only capable of achieving one of these objectives: they can either retrieve a large amount of information of low relevance (recall), or a small amount of very relevant data (precision). But the quality of matches is not the only issue—in many TM systems, performance degrades considerably as translation memories increase in size. Déjà Vu X's unique scan techniques—which even allow for the use of wildcards—and storage systems allow it to attain the perfect balance, displaying every useful translation available in a fraction of a second, whatever the size of your translation memory.
Assemble: Cutting-edge translation technology
Déjà Vu X's unique assemble feature combines example-based machine translation ideas with proprietary research. It takes advantage of the fact that, while sentence repetition may be scarce in many types of texts, repetition of smaller fragments is much more common.
Scan offers translations for sentences that match the complete sentence currently being translated; therefore, it won't display anything if it cannot find a good enough candidate. Assemble, on the other hand, will examine the associated databases more closely, taking relevant fragments, or sentences with a similar structure, to put together a translation from material that is embedded in other sentences and is therefore not available to scan.
The assemble function is available on an individual sentence basis, as a batch feature for the whole project, or as an automatic feature for each new sentence that is being translated.
Pretranslate: It works while you're on your coffee break
Both scan and assemble are extremely useful functions that considerably reduce the workload during the actual translation. In the case of very repetitive texts, though—such as documentation accompanying product upgrades or product catalogs—it is preferable to automatically process the project and insert the appropriate translations, thereby saving a significant amount of time and effort. This highly configurable batch process is called pretranslate. It involves Déjà Vu X examining the text that is to be translated and scanning the translation memory for matches for each sentence.
Déjà Vu will select the best translation—either an (appropriately indicated) guaranteed exact, exact or fuzzy match—for each sentence, and insert it in the corresponding location, leaving the translator with the task of reviewing and/or editing the translation rather than starting from scratch.
Propagate: Spread the word
Long and repetitive translation projects can pose several problems: it's not only unproductive to have to translate recurring sentences, but the translator runs the risk of translating them inconsistently.
Déjà Vu X helps to prevent these problems through the use of the propagate function. Whenever a sentence is translated, Déjà Vu can check the rest of the project for identical sentences, and automatically insert the same translation. If Déjà Vu X finds similar sentences, it will prompt for confirmation to insert a self-repaired fuzzy match. Just to make sure that the translation is indeed correct for that particular context, the propagated sentences will be indicated as such so that the translator can review them accordingly.
As with most other tasks in Déjà Vu X, propagate can be performed individually or integrated into the translation process as an automated feature.
AutoSearch: All the information at your fingertips
Content from your databases—translation memories, terminology databases, and the lexicon—can be accessed and used in various ways: from pretranslation to scan and assemble. But what about just viewing it as you translate? Déjà Vu X allows you to do just that with a feature called AutoSearch. Sentences, segments, and terms can be displayed with all the necessary context information and intelligently sorted according to relevance by Déjà Vu X. These windows will give you direct access to the displayed data for editing. And you'll be able to see where it differs from the text you are trying to translate.
And because this is a Déjà Vu X feature, it's naturally completely customizable. You can choose to have AutoSearch display only exact matches or all matches that it can find. You can decide whether or not complete sentences should be included in the AutoSearch window or whether they should be displayed in a separate window. Or you can decide not to use this feature at all. It's your choice.
Project management: Muster the troops
Coordinating large, multi-file, and multilingual translation projects can be a daunting task. The complexities of distributing work among translators, especially if working with both in-house and freelance translators, can often take up most of a project manager's time—time that would be better spent on monitoring the progress of the translation and providing support to the translators.
Fortunately, Déjà Vu's powerful project management features can significantly ease the worries of the project manager.
With Déjà Vu X, project managers can create and distribute satellite or Pack & Go projects, which are small, single-language files that upon completion can be reimported into the main multi-language project. Since they do not contain the formatting, image, and layout information that can make DTP format projects very large, they can be easily transmitted over the Internet, the company network, or other electronic media.
Along with satellites, project managers can also distribute smaller subsets of their databases, which contain only those sentences and terms that are relevant to the project. In this way they can provide reference material that is necessary to maintain consistency and simplify the translator's job, while at the same time assuring that one of their most valuable assets-the translation memories and terminology databases-are not being distributed too freely.
For groups of translators working together on a network, project managers can assign individual security access rights to each translator, allowing multiple translators to work on the same target language. Managers can also lock specific sentences or sections of a project to prevent translators from working on portions assigned to other colleagues. This allows managers to send complete satellite projects to translators, giving them full context information and at the same time ensuring that they only translate what is assigned to them. If it is necessary to supply additional information to translators, project managers can utilize the Comment function to add comments to each row (or to answer questions that were posed by translators using this function).
In other cases, the project manager may decide that certain translations should not be added to the translation memory until they have been validated, and can ask Déjà Vu to specifically ignore those sentences when sending information to the database.
One of Déjà Vu X's important project management features is the ability to analyze projects by counting words and characters per file or per project, generating an overview of the project's internal repetition, and finding out how many and what kind of matches there will be in the associated translation memories.
Déjà Vu X has extended the usefulness of pretranslation, the process of leveraging the database content to insert existing translations into a project. While Déjà Vu X still inserts fuzzy and exact translations, it also checks the surrounding sentences of each exact match and "guarantees" matches that are found to be identical, thus greatly reducing the need for additional proofing of the guaranteed exact matches.
And Déjà Vu X's extensive quality assurance features allow project managers to test the quality of the finished project—even if they don't speak the language!
Quality assurance: Be the best you can be
In addition to providing terminology management and glossary creation tools, Déjà Vu provides users with essential quality assurance functions that can check for inconsistent use of terminology, numerals, and coding, and can optionally tell them whether specific sentences or single terms have been translated in more than one way. Using these features, translators have much greater control over the quality of their translations, and project managers can make sure that their guidelines are being followed and detect and correct any possible oversights.
Integrated environment: You design the way you work
Déjà Vu X is not only a translation tool, but a highly customizable translation environment. It is designed to simplify the whole translation and translation management process.
Every task within Déjà Vu X is performed from and within the same user interface, no matter whether it is translating files of any of the supported formats, managing the databases, or building an SGML/XML filter.
Because we know that users have different preferences and different sets of tasks, we also designed the Déjà Vu X interface to be completely customizable, with user-definable menu bars and keyboard shortcuts, redefinable color schemes, freely placeable tools windows, and even the option of emulating the Déjà Vu 3 environment that so many of our users appreciate.
During the translation, all relevant memory and terminology database records are displayed in real-time (actually, no time) in AutoSearch window(s), and they can either be automatically inserted into your translation or added with a single keystroke. Any displayed database entry can also be modified or even deleted from within the AutoSearch display.
We've also accommodated those users who have grown accustomed to features from other tools, such as word processors:
- Our AutoCorrectAutoText features are similar to the corresponding Microsoft Word features. Often-used long words or expressions, or often misspelled words, can be added as global or project-specific AutoText/AutoCorrect entries that Déjà Vu X expands as they are typed. To maximize efficiency, it's also possible to import all of Word's AutoText and AutoCorrect definitions for each available language.
- Déjà Vu X includes spell checking facilities for 14 major languages. If you work in a language that is not supported, you can link into the Word spellchecker of that language, if you have it installed.
- Most of the common Windows commands and shortcuts for copying, cutting, pasting, searching, replacing, or selecting text have been implemented, and even unlimited undo and redo features are supported.
Multiple file format support: Everything under one roof
Déjà Vu X can build and handle projects for Microsoft Word, Rich Text Format, Windows Help, Microsoft Excel, Micosoft PowerPoint, Microsoft Access, OpenOffice/StarOffice, Adobe FrameMaker MIF, Adobe InDesign, Adobe PageMaker tagged text, QuarkXPress XTG, Interleaf ASCII, HTML, ASP/ASP.NET, PHP, JavaScript, VBScript, HTML Help, SGML, XML, RC, C/C++/Java, Java properties, IBM TM untranslated segments, Trados WorkBench documents, TradosTag TTX, Trados TagEditor BIF, Trados TagEditor TMX, GNU PO and POT files, and plain text files. The Workgroup edition also provides support for EBU subtitle files and ODBC-compliant databases.
Despite the wide variety of file formats, Déjà Vu X always displays the translatable contents (and only that!) in the same format, in project files with as many file types as it supports, and allows for the complete leverage of the content in your terminology and translation memories between the different formats.
For translators, this means that there's no training required to be proficient on a number of word processing, desktop publishing, software, and web development tools. And for project managers, it takes away any worries that the coding of their files may be corrupted.
Batch processing: The power boost
Batch processing is at the very core of what Déjà Vu X is about, and it shows in almost every respect:
- There is no limit to how many languages a project, translation memory, or terminology database can include, and all languages within a project can be pretranslated, translated, imported, or exported simultaneously.
- There is no limit to the number of files and supported file formats within one project, and all of them can be processed simultaneously or on a file-by-file basis.
- There is no limit to the number of users who can simultaneously access a project, a translation memory, or a terminology database.
Terminology management: Mind your words
Déjà Vu X's terminology databases are not mere glossaries of source and target terms. Instead, they follow a complex and completely configurable database concept that allows for as many attributes with as many categories as the user needs to properly define his or her terminology. In fact, the database concept that Déjà Vu X uses for its terminology management not only allows for the traditional bilingual translation relation between terms, but also monolingual ones such as synonyms or antonyms.
Database maintenance: Trim down and build up
All Déjà Vu X translation memories and terminology databases are multilingual. This not only means that there can be one source and several target languages, but the source language of one database can become a target for another project!
Déjà Vu X comes with database maintenance utilities that add an extra security level to your databases' integrity. Because Déjà Vu X automatically saves your work for you, there is usually no need to take any special precautions. However, after a hardware or power failure the database may be left in an unusable state. In this case, it might be necessary to use the database maintenance utilities to repair or partially rebuild the database.
Because Déjà Vu X employs the industry-standard Jet database engine, there are a variety of ways to communicate with your databases. Déjà Vu X provides a set of pre-defined tools that allow for database-wide search and replace operations, import and export of subsets, and filtering and cleaning the databases. But the possibilities for communicating with the databases do not end here. Users with a knowledge of SQL (Structured Query Language) can use that language to filter records, effect batch updates, or perform mass changes to their heart's content.
Déjà Vu X also includes a large number of import and export filters. These allow you to import glossaries and translation memories from a wide variety of sources, to export subsets of databases (this is useful when there are translations belonging to two different sectors or clients that use different target terminology), or to convert databases to other formats in order to exchange them with other translators or send them to your clients.
Déjà Vu X includes filters for plain text files (tab delimited, CSV, or character delimited), Access, Excel, IBM TM databases, Trados WB databases, the industry-standard TMX memory exchange format, and, in the Workgroup edition, for ODBC-compliant databases.
Alignment: A wealth of words
You may not realize it, but you are probably sitting on a gold mine. If you have previous translations, you can use the integrated alignment tool to identify sentence translations in those texts and add them to the translation memory, instantly gaining a wealth of useful material that will increase Déjà Vu's effectiveness and greatly enhance your productivity. While Déjà Vu X's Basic Aligner provides a user-friendly interface in which the alignment of single file pairs can be adjusted, the Translation Memory Builder is able to intelligently recognize matching patterns across languages and a large number of file pairs and perform an alignment that goes significantly beyond the mere splitting and matching of text in sentences.
Project lexicon: Set the terms
One of the challenging aspects of quality assurance in the translation industry is the enforcement of a consistent use of terminology for single translators or translation teams. The creation and use of glossaries is of paramount importance to this effort. Unfortunately, manually creating appropriate standardized glossaries is a time-consuming task.
Déjà Vu X solves this problem by allowing the generation of a lexicon—a project-specific index of words and phrases. Déjà Vu X will analyze the project and automatically extract words and short phrases from it. By displaying them in order of frequency, important terms and expressions are easily spotted, while non-relevant ones can be eliminated. Déjà Vu X also provides tools to automatically translate the lexicons, thus instantly generating terminology databases from existing projects. The final translated lexicons can be easily validated, making the implantation of standard terminology a feasible task.
Project lexicons make it easy for the project manager to distribute a unified glossary to all translation team members in a variety of formats or within Déjà Vu X. And they speed up the translation process for the translator. Many of Déjà Vu's translation features, such as pretranslate, assemble, or AutoSearch, make use of information in the lexicon, reducing the amount of work needed to translate texts. Since translators spend less time translating, they can devote more time to ensuring the correctness and consistency of their work, and hence increase the quality of their output.
View Cart
