CONTENTS Introduction ............................................................................................................................................................ 3 1 Overall Design Aspects ........................................................................................................................................ 5 Choose appropriate authoring software ........................................................................................................ 5 Design appropriate templates ........................................................................................................................ 6 Choose fonts wisely ........................................................................................................................................ 6 Use linked graphics ......................................................................................................................................... 6 Techniques for managing text expansion ....................................................................................................... 7 How to Improve the efficiency of translation tools .......................................................................................... 7 Misuse of carriage returns .............................................................................................................................. 8 Placement of index markers ........................................................................................................................... 8 Manual hyphenation ...................................................................................................................................... 9 Conditional text .............................................................................................................................................. 9 2 Writing Documentation .................................................................................................................................... 10 Keeping content simple ................................................................................................................................ 10 Be consistent ................................................................................................................................................ 11 Ensure that content is culturally acceptable ................................................................................................ 12 Guidelines for the document translation process .......................................................................................... 13 Translating Adobe FrameMaker files ............................................................................................................ 14 Translating Microsoft Word files .................................................................................................................. 15 collateral and marketing material ................................................................................................................ 16 Collateral authoring tools ............................................................................................................................. 18 3 Writing Online Help .......................................................................................................................................... 22 Guidelines for the Help translation process .................................................................................................... 23 XML guidelines .............................................................................................................................................. 23 Translating and localizing body text ............................................................................................................. 24 Guidelines for JavaHelp ................................................................................................................................ 25 What translation involves ............................................................................................................................. 26
Page | 1
4 Preparing for Web Site Translation ................................................................................................................... 28 Global design issues ...................................................................................................................................... 28 Style and layout ............................................................................................................................................ 30 Electronic commerce .................................................................................................................................... 30 Ongoing website management ..................................................................................................................... 31 5 Brand name issues ............................................................................................................................................. 33 6 When things go wrong ....................................................................................................................................... 34
Page | 2
INTRODUCTION This guide has been developed by the technical staff of Academy Translations to give technical writers a comprehensive overview of how to structure and write documents in such a way that it will be easier to translate them. Some basic modifications to typical technical writing will have a marked positive impact on the speed and cost of future localization and translation. In brief, use controlled terminology, use a simple unambiguous style, allow for extra space in tables and illustration and avoid country‐specific references.
“Localization is the process of making a product linguistically and culturally appropriate to the country, region and language where it will be used.”
Controlled terminology Uniform terminology will minimise translation effort, maximise consistency and accuracy of the translated document and allow for the most efficient use of translation support tools. Terminology is best controlled by creating a glossary of terms relating to the product and actions required by the software (for example, click, select, and so on), and applying this consistently throughout each document and between documents relating to a product or a suite of products. In particular, electronic and printed documentation should use the same set of terms. Simple, unambiguous writing style Controlling writing style is also worthwhile. This includes avoidance of jargon, long sentences, cultural references, acronyms and abbreviations. Use concise sentences, preferably in the active voice. All dates need to include a written name of the month as the date and month order varies between countries, and abbreviated dates can be confusing. Controlled language These issues have been addressed by the development of “controlled English” or easy English standards for writing documentation. This is a modification of English that uses a set of standard terms and phrases and simplified grammar. For example, there is no use of the passive tense, a maximum number of words per sentence and so on. There are also lists of words and phrases to be avoided. Page | 3
Controlled language reduces ambiguity and promotes readability. It makes writing initially more difficult, but translation is much simpler, and allows for optimal use of translation software. It can reduce translation time by as much as 25%. There are several commercial software tools for writing controlled English. For further information about controlled language, consider these sources:
•
Information on Controlled English – KANT (Knowledge‐based Accurate Natural language Translation)
•
Controlled language software ‐‐ MAXIit by SMART Communications – www.smartny.com
Page | 4
1 OVERALL DESIGN ASPECTS
• • • • •
Choose appropriate authoring software Design appropriate templates Ensure that chosen fonts are localizable into the target language Link graphics rather than embedding in the document Techniques for managing text expansion
CHOOSE APPROPRIATE AUTHORING SOFTWARE It is wise to be aware of translation issues when choosing word processing software to produce documentation. The same application should be used for all language versions of the document, but to do so, the application has to support all the target languages. Verify with the translation vendor that your software is compatible with computer aided translation tools they use. Ideally the chosen software should: • • • • •
Suit the document type Support all target languages Support the right output formats Be compatible with computer aided translation tools Support single‐sourcing if required (one source for Print, Online Help and Web)
From experience we recommend to avoid using low‐end word‐processing software (such as MS‐Word). Whilst there would be an initial cost saving by using such software because all technical writers are familiar with it to some degree, the added cost, time and frustration of ongoing formatting problems in other languages makes it an uneconomical choice. We recommend using authoring software such as Adobe FrameMaker (structured file format if possible) or HelpStudio (www.innovasys.com) or similar. HelpStudio, for example, allows you to export the translatable text in XML format. The translation vendor will use a Translation Memory (TM) tool to process and translate the files, and return the translated files in XML format for you to import. All formatting will remain intact.
Page | 5
If done correctly by your translation vendor there is no need for cutting, pasting and reformatting of each language. The Table of Contents (TOC) and Index files will also be automatically updated because all marker text was extracted and translated together with the visible text. Whilst an Index at the end of a manual is common practice for English and European‐ language manuals, this is not always so in languages such as Chinese, Japanese, Korean and Arabic as these scripts do not allow alphabetical listing of key words. Your regional sales manager in these countries, or your translation vendor, should be consulted to advise what the convention is for your particular document. As stated before, none or only minimal formatting adjustments should be necessary for most languages after translation. The only exception is the Thai language, which requires the manual adjustment of line endings by the translator and will therefore cost a little more.
DESIGN APPROPRIATE TEMPLATES If low‐end word‐processing software has been selected to start with, the page layout needs to be chosen carefully to minimize the cost of the post‐translation formatting. This can sometimes represent up to 40% of the total translation cost, and attending to layout issues can affect price significantly. Use a ‘universal’ page format that contains enough room for text expansion in tables, paragraphs and leading. Generous white space in the English version (which will be pleasing to the eye), will accommodate text expansion in the foreign language versions. A simple and cost‐effective template design approach is to use automatic rather than manual generation of tables of contents and indexes.
CHOOSE FONTS WISELY Choose fonts that also include all the appropriate characters for most target languages. Doing so will preserve the document’s look‐and‐feel.
USE LINKED GRAPHICS Avoid embedding graphics in documents but link them instead. They will be easier to modify for a particular locale, if necessary. Page | 6
Ensure that any captions are a part of the document, not the graphic itself. This will minimize updates to graphics by translators.
TECHNIQUES FOR MANAGING TEXT EXPANSION
• •
Prepare English document template with additional space Be aware of average expansion and contraction rates
When preparing the layout of documentation, aim to account for approximately 35% expansion of text in translation. To do so, the most efficient way is to prepare a document template for English that has some reserved space. Some helpful techniques include: -
Having a slightly shorter text body in the English version, thus reserving additional white space at the bottom of the page
-
Always setting up the footer space so that it can include at least two lines of text
-
Making columns in tables slightly wider than necessary
-
Avoiding a justified paragraph style, especially when no hyphenation is to be used. Justification in languages with longer word length can produce unpleasantly large white spaces between words
-
Avoiding deeply nested lists as these tend to reduce column width
If these guidelines are observed, your translation vendor will be able to import the English version into a translation memory (TM) tool such as Atril’s Déjà Vu X (DVX) or SDL Trados, translate it then export a formatted translation using the original English layout.
HOW TO IMPROVE THE EFFICIENCY OF TRANSLATION TOOLS • • • • •
Do not place manual carriage returns at the end of each line of running text Do not place index markers inside sentences or words Use ‘fields’ with caution Avoid manual hyphenation Avoid the pitfalls of conditional text
Page | 7
MISUSE OF CARRIAGE RETURNS Instead of pressing Enter at the end of line, allow the authoring software to insert a soft line break that can be adjusted automatically to accommodate changes in text. Manual carriage returns will be propagated to all translations by the TM software of the translation vendor and will result in extra formatting cost.
PLACEMENT OF INDEX MARKERS Index markers are used in authoring packages to automatically generate indexes. Each index term has a corresponding index marker on the page where the associated text appears. It is good practice to insert these markers before or after the relevant heading and not in the middle of sentences so that these can be easily identified for translation by your translation vendor’s TM software.
USE MS‐WORD FIELDS WITH CAUTION Fields in MS‐Word are a great time‐saving feature to propagate recurring words and phrases throughout the document. This is true provided that the document will never be translated. However, when a translation of such a document is attempted it becomes clear very quickly that this feature must be used with the utmost caution. Consider this real‐life example: The author of a manual used Fields to refer to a menu option called “Save records”: Save records – This feature is used to save your records. This was translated into Dutch as: Records opslaan – Met deze functie kunt u records opslaan. However, now the translation of “Save records” was set in concrete. It was automatically applied throughout the rest of the manual wherever “Save records” appeared in English. Now consider this sentence: To save records, use the Save Records option in the File menu. This would normally be translated into Dutch as: Om records op te slaan, gebruikt u de optie Records opslaan in het menu ‘Bestand’. However, the Fields option forced the translation in this way: Om records opslaan, gebruikt u de optie Records opslaan in het menu ‘Bestand’. Page | 8
The problem is obvious. The Field caused an automatic wrong Dutch translation simply because Dutch grammar doesn’t work like English grammar does. As a result the entire manual had to be edited and the use of Fields restricted to sections where it couldn’t cause any inadvertent mischief.
MANUAL HYPHENATION Authoring packages provide the option to automatically hyphenate when necessary. This is preferable to manual hyphenation because hyphenation rules change between languages, and automatic hyphenation deals with this.
CONDITIONAL TEXT •
Conditional text has the advantage of making it easier to re‐purpose your text, or provide different variants of a document to suit different contexts (for example User Manual versus Online Help). However, be aware that this can add an additional layer of complexity to the translation task if not done properly.
•
The point of concern is to tag whole paragraphs as conditional rather than narrowing the scope to individual sentences or words. This provides more context to the translator and avoids problems due to differences in grammar between English and the target language. By keeping conditional text at the paragraph level, you give the translator adequate ‘space’ or context to work with.
•
If done properly, conditional text can actually reduce the translation effort, because text that is common to a number of documents only needs to be translated once. Remember, however, that each item of documentation should still be proofed independently.
Page | 9
2 WRITING DOCUMENTATION
When developing a software package, remember that the documentation is as much a part of the product as the code. As soon as a software glossary or a preliminary build of the localized software is available, translation of the online help and documentation can start.
“Keep content simple, consistent and culturally neutral.”
For product documentation to be globalized and localizable, the content (text, art or multimedia) should be: • • •
Written as simply as possible Written as consistently as possible Culturally acceptable and inoffensive to an international audience
KEEPING CONTENT SIMPLE
The following guidelines help to keep your content simple: •
Be concise
This assists speakers of your own language and also makes it easier to translate. •
Use active voice
This shortens sentences and is more familiar to international readers. •
Avoid long‐winded expressions
For example, instead of writing ‘in order to’ just write ‘to’. •
Use simple verb forms
For example, write ‘select’ instead of ‘make a selection’.
Page | 10
BE CONSISTENT • • • • • • • •
Use words with a precise meaning Use the same term to express a specific meaning Include ‘optional’ words Use consistent capitalization Use more punctuation, not less Avoid alphabetical lists that cannot be generated automatically Do not use nouns as verbs Avoid words that have opposite meanings
The following guidelines will increase comprehensibility and make localization easier: •
Use words with a precise meaning Unambiguous terminology aids understanding. For example, use ‘Install the application’ rather than ‘Set up the application’.
•
Use the same term to express a specific meaning For example, do not use ‘remove’ and ‘delete’ interchangeably.
•
Include ‘optional’ words There is a tendency to omit words in sentences, which increases the chance of misunderstanding. For example, ‘You can change the sysdb file using the BRG utility.’ This can mean either: ‘You can change the sysdb file that uses the BRG utility.’ or ‘You can change the sysdb file by using the BRG utility.’ Also, avoid omitting articles such as ‘a’, ‘an’, and ‘the’ to add clarity.
•
Use consistent capitalization This will help to standardize terminology for the localizer.
•
Use more punctuation, not less This reduces confusion for both the reader and the translator by breaking complex sentences into parts for easier understanding.
•
Avoid alphabetical lists that cannot be generated automatically
Page | 11
Unless lists (such as commands, keyboard shortcuts and glossaries) can be generated automatically, compiling these alphabetically for each language will be very time‐consuming. •
Do not use nouns as verbs For example, ‘This function gives an analysis of the problem and offers a solution.’ The nouns ‘analysis’ and ‘solution’ convey most of the meaning in this sentence, while the verbs ‘gives’ and ‘offers’ are practically meaningless. A better sentence would be: ‘This function analyzes the problem and solves it’.
•
Avoid words that have opposite meanings For example, ‘sanction’ can mean either ‘to approve’ or ‘to punish’.
ENSURE THAT CONTENT IS CULTURALLY ACCEPTABLE • • • • • •
Remove text that may be culturally sensitive Avoid culture‐specific metaphors Avoid humor Avoid ambiguous words and jargon Spell out acronyms Avoid clip art
Use the following guidelines to ensure that content accounts for the wide spectrum of cultural differences among international readers: •
Remove text that may be culturally sensitive to ensure neutral, culture‐ independent content Do not use slang or colloquial expressions; readers may not be able to distinguish the idiom from the literal meaning. Idiomatic expressions are also difficult, sometimes impossible, to translate.
•
Avoid culture‐specific metaphors Metaphors can have cultural associations that are not relevant outside of your country.
•
Avoid humour
Page | 12
Even within your own country, people have differing views about what is funny. Humour does not add clarity to international content, and is often idiomatic. •
Avoid ambiguous words and jargon Jargon is technical terminology that has been modified by those in a particular profession. It assumes both technical and local knowledge and is idiomatic to a particular professional ‘culture’. Jargon is a ‘short‐cut’ but only for those few who know it, and is inappropriate for international content.
•
Spell out a term’s acronym on the first occurrence, followed by the acronym in parentheses Subsequent usage can be confined to the use of the acronym. In non‐linear documentation (such as on‐line help), where material is unlikely to be read sequentially, a term should be spelled out on the first occurrence for each topic in which it appears.
•
Avoid clip art Clip art is usually inappropriate for international customers. For example, clip art should not be used to represent currency, which has different graphical representations from country to country. Likewise, icons are rarely culturally‐ neutral (such as mailboxes, road signs and forks and knives). Wherever such items are used, the localizer must replace the graphic with an equivalent; in most cases, none exists.
GUIDELINES FOR THE DOCUMENT TRANSLATION PROCESS • • • •
Generated files are not translated Hidden text may or may not be translated Provide a formatting style guide Provide a language style guide & terminology glossary
•
Generated files These include such items as the table of contents and index.
•
Hidden text Page | 13
Hidden text in the document may or may not need to be translated, depending on the file type. This text may therefore need to be audited and the relevant items for translation identified. •
Formatting
Provide the localization vendor with a formatting style guide (e.g. all references to GUI buttons should be in boldface). •
Terminology
Provide your translation company with a language style guide and terminology glossary.
TRANSLATING ADOBE FRAMEMAKER FILES •
Change bars
Ensure that change bars are switched off. •
Fonts and images
Ensure that there are no missing fonts or images before issuing the files. •
Table of contents, index and cross‐references
Ensure that the table of contents and index generate correctly and that there are no unresolved cross‐references. •
Headers and footers
Headers and footers often contain variables that automatically insert items such as chapter numbers or section titles. These do not require translation. •
Markers
Headers and chapter titles are often included in markers. Index entries also use markers. Text contained in markers will need to be translated. If your translation company uses the latest TM software, these markers will be extracted and translated automatically. •
Cross‐references
Page | 14
Cross references are automatically updated and do not need to be translated. However text contained in the cross reference definition will need to be translated (such as the words “on page” which precede the page number). •
Conditional text
Conditional text is text that appears or is hidden depending on certain conditions. This enables a document to be repurposed, depending on context. For example, platform‐specific information could be hidden or revealed in the case of a multi‐ platform product. Confirm which text is to be translated, though in general all text should be translated.
TRANSLATING MICROSOFT WORD FILES •
Styles
Most TM tools support Microsoft Word by default. However, ensure that styles have been set up and no manual style overrides are used. After translation with a TM tool, the original style sheet may need to be attached. •
Revisions
Ensure that files contain no revision marks. •
Headers and footers
Headers and footers need to be translated. •
Index entries
Indexes tend to be generated from index entries or markers, which have been inserted in the main body of the text. The generated index is never translated, as the translation will be lost the moment it is regenerated. The index entries and markers that are used to generate the index are translated instead. As discussed earlier, indexes are only useful in languages that can be sorted alphabetically. They are often omitted in languages such as Japanese, Chinese, Korean, Thai and Arabic. Please consult your local sales office or your translation vendor for country‐specific advice. •
Cross‐references
Page | 15
Cross references are automatically updated and do not need to be translated. However text apart from the heading text and page number will need to be translated (such as the words “on page” which precede the page number). •
Fields
Use this feature with the utmost caution. See page 9 for details.
COLLATERAL AND MARKETING MATERIAL • • • •
Avoid text in graphics Be mindful of cultural differences Allow space for text expansion Avoid metaphors and complex examples
Collateral material is often market‐oriented (e.g. packaging and promotional material) or country‐specific (e.g. specifics of product registration). The following points should be taken into consideration: • •
Do not include text in graphics Be aware of culture‐specific responses to images, for example: • Keyboards, and therefore their images, differ between countries • Female faces or figures may be inappropriate in Muslim countries • Colors have specific meaning in different cultures
•
Allow space for text expansion
•
Aim for clear communication without metaphors or complex examples
•
Be aware of cultural issues in advertising
Types of collateral material • • • • • • • •
CD labels Product packaging Reference cards and Order forms Marketing material Registration cards Installation procedures Promotional materials Warranty cards Page | 16
Characteristics of collateral material • • • • •
Many images Complex layout Text minimal but often very market‐oriented Creates the first impression of the product May contain country‐specific technical or legal material
Collateral material usually contains little text, but often entails complex layout and numerous images. Do review the images for cultural appropriateness. Some collateral material may not be suitable for TM tools because: • • •
There is too little text to justify the effort of importing into a TM Translators need to see the text in context Rewriting of content may be required
An example for this would be a CD label that may come as an .EPS or .AI file for translation. Ensure that the translation vendor can use the software tool that the material was created with, and assume that there will be a post‐translation layout and desktop publishing phase before publishing. Other collateral material may well be suitable for TM tools. This may include marketing and promotional material and installation instructions that contain quite a substantial amount of text and were formatted in packages like Adobe InDesign or QuarkXPress. A good translation vendor will have the technical know‐how and software to extract the translatable text from these file formats automatically, perform translations using a TM tool and then export the translations in the original file format, fully formatted. Using this method, the DTP effort is greatly reduced, all translations are reusable because they are in memory, and the completion time is much faster than using the old copy & paste method. This saves you a substantial amount of money and time over the life of the product. Guidelines for translating collateral material • • •
Clear, straightforward text is the easiest to translate Minimize marketing slogans – they can be difficult to translate effectively Be aware of culture specificity of images Page | 17
• •
Do not include text in graphics Allow space for text expansion
Collateral material is more culture‐specific than the rest of the product. Marketing material is often geared to a specific market and marketing slogans are notoriously hard to translate well. Attempting to translate marketing slogans is not recommended. Instead, engage a copywriter from the target country to write a new one that is appropriate for the country or region. In fact, the more clever the marketing slogans and marketing images, the more difficult they are to adapt to international markets. Symbols and metaphors are culture‐specific and need to be replaced with the relevant equivalent for each locale. In many cases, there are no equivalents due to idiomatic differences. The materials that translate easily are straightforward product descriptions that do not use catchy phrases and metaphors.
Beware of slogans! “For example, ‘saving effort’ is not a positive feature in Japan, where effort is one of the main virtues.”
Legal text (warranty and disclaimers) should, after translation, be checked and edited by a legally qualified person in the target country to avoid infringement of local laws and statutes. It is worthwhile spending some time checking the appropriateness of graphics at this point, so that these do not need to be changed when the software is localized. Also, it is important to include any text in a separate layer of the graphics, so that it can be easily substituted with the translation. Collateral material often has complex layout, and it is particularly important to remember the additional space requirements of translated text. Be aware that some legal or technical details may change between countries. Try to ensure that such details do not appear in multiple locations, so that only one document needs to be updated by technical or legal specialists in a given location.
COLLATERAL AUTHORING TOOLS Popular tools used to develop collateral include: Page | 18
• • •
QuarkXPress – for packaging, quick reference cards and documents of one or two pages. Adobe InDesign – for multi‐page documents, such as introductory guides Adobe PageMaker –for multi‐page documents
What follows are guidelines for localizing material authored with these tools. Preparing QuarkXPress files • • • • • •
Save previews of images. Ensure images that contain text are issued to the translator. Save Quark files in multilingual format to assist translation (QuarkXPress 4 or later). Resize text boxes to allow for text expansion (preferred), or unlock them to enable expansion. Do not use the Text to Box command (QuarkXPress 4 or later). Index entries will need to be translated (QuarkXPress 4 or later).
QuarkXPress is a powerful tool with high‐end graphic, typographic and printing capabilities. Quark is also available in a multilingual version, called QuarkXPress Passport, the interface of which can be run in different languages and enables hyphenation and spell checking in many languages. Documents created with QuarkXPress 4 and later can be saved in multilingual format for use by QuarkXPress Passport. It is not necessary for you to invest in the Passport version of QuarkXPress, but saving files in multilingual format will assist the translation process. Guidelines •
Save a preview of an image when saving a document. This will ensure that the image is displayed in the document even if the original image file is absent. However, if an image contains translatable or localizable text, ensure that the image file is included.
•
Save Quark files in multilingual format to assist the translation process.
•
Resize text boxes to allow for text expansion (preferred), or unlock them to enable the translator to expand them as necessary. Resizing in advance means layout will not change significantly as a result of translation.
Page | 19
•
Do not use the Text to Box command, which saves text as outlines (QuarkXPress 4 or later). Affected text cannot be edited or restored.
•
Index entries will need to be translated (QuarkXPress 4 or later).
Preparing Adobe InDesign files What to send to the translation vendor: • • •
The INDD file or files All images files All fonts that occur in the document
Allow sufficient room for text expansion in stories and tables. Adobe InDesign is a page layout tool for medium‐sized documents. It integrates with Adobe InCopy, a professional writing and editing program that fosters a collaborative workflow with authors and graphic designers. However, these products do not include tools for the translation process. Creative work is organised into ‘stories’. A story is a useful analogy in the design world; it is a collection of text objects, such as those contained in a magazine article. Typically a story is a text box. Translation of an InDesign file involves: 1. Exporting all stories in the InDesign file to a generic format such as Rich Text Format (RTF) or HTML 2. Having the generic file translated by a third party 3. Importing the translated text into the InDesign file Unfortunately, neither InDesign nor InCopy provide a means to export all ‘stories’ at once. This means that exporting material is a piecemeal process that is prone to error through the omission of text. Importing the translation into InDesign is similarly problematic, with the additional difficulties in retaining all the original formatting. Nevertheless, there are plug‐ins by third parties available to handle this. One example is ECM Engineering’s InDesign filter (http://www.ecm‐e.de/). A good translation vendor will have this plug‐in, which makes the translation of InDesign files very quick and less labour‐intensive in terms of reformatting, thus saving you time and cost. The added advantage is that TM tools can be used in this process.
Page | 20
If you allow sufficient room for text expansion in text boxes, and in particular in tables, the translation vendor will take care of the rest and should deliver a translated and fully formatted InDesign file to you. A translation memory will have been created along the way, which will save you costs in translating an updated document at a later stage. Localizing Adobe PageMaker files • •
•
Ensure images that contain text are issued to the translator. If a TM is to be used: – all ‘stories’ should be contained in a single text flow, and – all paragraphs should have a tag. Index markers will need to be translated.
Adobe PageMaker is a page layout tool for medium‐sized documents. PageMaker enables you to generate a table of contents and an index that spans several documents.
Page | 21
3 WRITING ONLINE HELP
Online help is often the largest translation component of a localization project. Online help files and online manuals have largely replaced traditional printed documentation for a number of reasons, including the following: • • • • •
Context‐sensitive help gives users instant access to documentation from the software. Hypertext and index keywords provide fast navigation to detailed information Online help can include multimedia features such as animation, movies and sound. Updates can be distributed easily via the web. Online documentation provides significant savings in production costs.
Online help source files • • • •
Text and formatting Images Multimedia Document layout
Most online help is created from a set of source files, which are compiled into one central binary help file. Source files typically include: •
Text and formatting
For example, RTF files (WinHelp), HTML files (HTML Help) and XML/HTML files (JavaHelp) •
Images
For example, .bmp files (WinHelp) and .gif or .jpg files (HTML Help) •
Multimedia
For example, movies and sound effects •
Document layout Page | 22
For example, Cascading Style Sheet (CSS) files for HTML‐based help
GUIDELINES FOR THE HELP TRANSLATION PROCESS • • • •
Hidden text and coding are not translated Provide a formatting style guide Provide a language style guide & terminology glossary Hidden text and coding
Hidden text in the document will not be translated; neither will codes, such as references to screen shots, index markers, encoding and formatting tags. •
Formatting
Provide the localization vendor with a formatting style guide (e.g. all references to GUI buttons should be in boldface). •
Terminology
Provide the localization vendor with a language style guide and terminology glossary. Alternatively, the localization vendor can create a terminology list with translations of terms that occur frequently in the source text; you should approve this list before translation proceeds. A software user interface glossary is also highly recommended.
XML GUIDELINES eXtensible Markup Language (XML) is both a procedural and descriptive markup language. XML’s descriptive tagging enables the types of information in a document to be defined, and how the information interrelates. XML is a metalanguage: it enables you to create your own markup, defined in a Document Type Definition (DTD). Considerations: • •
Files requiring localization Translating and localizing the Header Page | 23
•
Translating and localizing body text
As with HTML, the procedural markup defines the layout of text (such as heading styles). XML was created as an alternative to HTML, which proved too unsophisticated for interactive web sites and web‐based applications. XML’s descriptive tagging means that the types of information contained in the document can be defined, and how these items of information interrelate. For example, a recipe would likely include several categories of information: a title, a description of the dish, ingredients, and a method. Each item of content is a tagged element of the XML document (for example ‘1 egg’ would be tagged as an ‘Ingredient’). Apart from being able to include both descriptive and procedural markup, XML is a metalanguage: it enables you to create your own markup. Tags used by an XML document are defined in a customized Document Type Definition (DTD). Although the XML document usually contains all of the translatable text, the associated DTD may also contain material for translation or localization, and therefore should also be issued for translation. Files that are likely to be included in an XML‐based online help system and require localization are as follows: • • • • • • •
Active Server Page file (.asp) Cascading Style Sheet (.css) Graphics (.gif, .jpg) HTML file (.html) JavaScript file (.js) XML file (.xml) XML stylesheet (.xsl)
TRANSLATING AND LOCALIZING BODY TEXT •
Text
Paragraphs are generally enclosed in markup that is defined in the DTD and describes the content, for example:
Jean Page | 24
Vioget French English <Specialty>Chemistry •
Graphics
Graphics are linked to the document in a number of ways, however the HTML
Comments
Comments are enclosed in tags, and are generally not translated. For example: ' If there is embedded code in the document, comments within the code are instead indicated by a double slash (//). •
Embedded code
XML documents can include embedded code, such as JavaScript. This code may be embedded in the XML document or contained in an external file that is referenced, for example:
To localize web‐based applications, as with regular software localizations, translators need to have access to the associated program resource files.
GUIDELINES FOR JAVAHELP JavaHelp is a platform‐independent help system for Java‐based applications that works with either normal or compressed HTML (.jar) files. Considerations: • •
Files requiring localization What localization involves – Help text – Help settings – Table of Contents & Index Page | 25
JavaHelp is a platform‐independent help system for Java‐based applications. It includes a viewer and works with both normal and compressed HTML files. The compressed form is the recommended format, and has a .jar extension. It encapsulates all HTML files and images into a single compressed file, similar to the .chm format used by Windows HTML Help. Source files for JavaHelp projects include HTML documents, .gif and .jpg images. JavaHelp is compiled using a project file with the extension .hs. This is an XML file that contains all settings for compilation. The project file is compiled by running the JAR command, which is available with the Java Development Kit (JDK). For further details, please browse to Java.sun.com. The HTML files in a JavaHelp project can be edited using any HTML editor, but they should ideally be edited using the tool that was used to create them. Typically the compiler is run through a macro shell application, such as WebWorks Publisher or RoboHelp HTML. Such tools provide additional authoring facilities. Localizable files that may be included in a JavaHelp localization kit are as follows: • • • • • • • • • •
Cascading Style Sheet (.css) Graphics (.gif, .jpg) HTML file (.html) Java file (.java) JavaHelp HelpSet file (.hs) JavaHelp window definition (.hsw) JavaScript file (.js) Properties file (.properties) XML file (.xml) XML stylesheet (.xsl)
WHAT TRANSLATION INVOLVES It involves translating the help text contained in the HTML source files, images and the XML files that contain the project settings, the table of contents, the keyword index, and file mappings. Each of these files has an associated DTD that defines markup and structure. These can be found in the JavaHelp development environment in \doc\spec\dtd. •
Help text
Page | 26
HTML files are either translated using a TM tool, or by editing them with the authoring tool used to create them. For more information, refer to ‘HTML Guidelines’. •
Help settings
The HelpSet (.hs) file is an XML file that contains all compilation settings for the JavaHelp project. It can be edited using either a text or XML editor. Parts of the .hs file that may need translation include the following. •
JavaHelp Title
The words between the
and tags are displayed in the help’s title bar and need to be translated. •
View definitions
The view definitions are contained within the
and tags the can contain labels for the table of contents, index and full‐text search. These will need translation. •
Table of Contents and Index files
The Table of Contents and index are each defined by an XML file (eg. IdeHelpIndex.xml and IdeHelpTOC.xml). Each of these contains lines of text with the text= attribute. In each instance of this attribute, text following text= is to be translated.
Page | 27
4 PREPARING FOR WEB SITE TRANSLATION
• • • •
Global design issues Style and layout Electronic commerce Management of a multi‐lingual website
Translating web sites is much more integrated with the process of localization and is often referred to as web site globalization. Globalization of a website includes: • • •
choice of software that supports multilingual hosting, preparing an appropriate multilingual website structure and navigation system, and allocating content to be either translated or rewritten for a given country.
Another important component is the design of an effective multilingual navigation system – one that can possibly track a user’s language preference through use of a cookie. This needs to be accessible on every page because of the potential to access the site through a search engine. Optimally, an internationalized web site: • •
Minimises the use of text within graphics Supports international characters through the use of Unicode
Some useful resources include: • •
World Wide Web consortium, which provides a Localization and Internalization web page: www.w3c.org/international Mark Bishop’s book on “How to Build a Successful International Web Site” – www.multilingualwebmaster.com
GLOBAL DESIGN ISSUES Page | 28
• • • •
An appropriate development platform and hosting environment An appropriate file structure Identifying material for translation and re‐writing Separation of text and code where possible
Multilingual web sites are sometimes the most easily accessible information about a product or a company. Their dynamic character, with hypertext, frequent updates and current local information make updates frequent and local information quite important. The right choice of hosting environment and site structure can determine the ease and cost of converting and managing the site in a multilingual form. Development platform and hosting environment One of the most fundamental decisions to be made is the choice of appropriate hosting environment and development platform. At the very least, multilingual scripts should be supported through the use of Unicode. However, there are tools that enable easy content creation, maintenance, and automatic translation management using a centralized text database architecture. Some of these tools include Webplexer and Global Mirror. Although many issues with website structure are addressed by these tools, there may be cases where a website development tool is not used. Key issues are briefly surveyed below. File Structure An issue to address at the outset of website planning is file structure. It is wise to create folders dedicated to each locale, each with its own subset of pages and images, each with its links adjusted to lead to pages within this language setting. In some cases, it may be easier to create separate websites for each language, in other instances, one website with duplicate pages for each language is more appropriate. Identifying material for translation While deciding on a file structure, it is worthwhile to decide which pages are to be translated and which need to be re‐written for each locale to reflect the relevant local data, such as contact and distribution details, warranty information and other country‐specific issues. Separation of text and code It is worthwhile to set up a database structure for the website, so that the text is separated from the code. This will increase the ease of translation and future updates, as the translator or website owner won’t have the added task of separating Page | 29
text and code and testing the website to ensure that no contamination of source code has taken place. However, keeping text in a separate file is not always possible or feasible. A translation vendor with the appropriate TM tools can quite easily handle HTML or XML files where code and text are mixed, without the danger of corrupting the code during translation.
STYLE AND LAYOUT
• • •
Use of style sheets Multilingual navigation system Appropriate use of graphics
Style sheets Formatting issues can affect the ease of translation and the quality of the final product. The best practice is the use of style sheets for formatting. This provides global control and avoids local code that can complicate the translation process. Be aware of the potential for text expansion as a result of translation, and account for this when preparing language‐specific style sheets. Multilingual navigation It is also important to design a multilingual navigation system that is accessible from each page. This caters for visitors accessing the web site from search engines. The names of languages should be written in the appropriate language and font, and optimally the user’s language choice should be tracked by a cookie. Use of graphics Images on a website often need to be localized for both content and text. It is wise to limit such images, and to keep them in separate image folders for each language for easy access.
ELECTRONIC COMMERCE
Page | 30
• • • • •
Foreign currencies, currency conversion Varying time, date and number formats Local addresses for shipping Shipping costs Awareness of local legal and commercial issues
Web sites that provide worldwide electronic commerce have their own set of requirements above and beyond basic localization issues. Although this is not a focus of this guide, a few of the key issues are highlighted. There is no question that being able to support worldwide purchases is essential, and yet many websites are incapable of dealing with customers from around the world. It is important to recognize the issues that need to be handled faultlessly for international commerce. These include: • • • • •
Dealing with foreign currencies and supporting currency conversion Supporting local time, date and number formats Dealing with local scripts (preferably through Unicode), so that names and addresses can be encoded even if written in a non‐Roman script Safe and effective credit card processing The ability to calculate shipping costs
ONGOING WEBSITE MANAGEMENT • • •
Frequent updates Use of translation tools Automatic multilingual update tools
Websites should be updated on a regular basis. They are much more dynamic than printed information, and users expect good websites to be up‐to‐date. As a minimum, they need to reflect current contact information. For larger websites and those in multiple languages, the use of tools for multilingual management is the recommended option. Such tools will flag updated text, and may even automatically initiate the required translations. Some of these tools include Webplexer and Global Mirror. If the use of such a tool is not feasible, you can minimise costs and translation time through an ongoing localization vendor relationship. The translation of any updates can be supported by previously translated text through use of a translation memory (TM) tool. In addition, a formal update plan that includes translation of updated text Page | 31
needs to be used to ensure that data is correct and up‐to‐date on the non‐English pages of the website.
Page | 32
5 BRAND NAME ISSUES
Brand names are a very special case. Don’t assume that just because the meaning of a name is obvious in your language and your country, that it will be recognized in other countries as well. Brand names must be ‘re‐engineered’ for another language and culture to have the meaning and impact desired by your marketing department. Failure to do so will most likely result in a failed product launch.
A real life story Apparently Coca‐Cola's initial transliteration of their name into Chinese produced a rendering with the meaning "bite the wax tadpole." When Coca‐Cola first entered the Chinese market in 1928, they had no official representation of their name in Mandarin. They needed to find four Chinese characters whose pronunciations approximated the sounds "ko‐ka‐ko‐la" without producing a nonsensical or adverse meaning when strung together as a written phrase. Written Chinese employs about 40,000 different characters, of which about 200 are pronounced with sounds that could be used in forming the name "ko‐ka‐ko‐la." While Coca‐Cola was searching for a satisfactory combination of symbols to represent their name, Chinese shopkeepers created signs that combined characters whose pronunciations formed the string "ko‐ka‐ko‐la," but they did so with no regard for the meanings of the written phrases these formed. The character for wax, pronounced "la," was used in many of these signs, resulting in strings that sounded like "ko‐ka‐ko‐la" when pronounced but conveyed nonsensical meanings such as "female horse fastened with wax," "wax‐flattened mare," or "bite the wax tadpole" when read. Coca‐Cola had to avoid using many of the 200 symbols available for forming "ko‐ka‐ ko‐la" because of their meanings, including all of the characters pronounced "la." They compromised by opting for the character lê, meaning "joy," and approximately Page | 33
pronounced as "ler." When transliterating the name 'Coca‐Cola' they finally settled on using the following characters:
This representation is literally translated as "to allow the mouth to be able to rejoice”. However, it acceptably represents the concept of "something palatable from which one receives pleasure”. It was the real thing, with no wax tadpoles or female horses, and Coca‐Cola registered it as its Chinese trademark in 1928.
6 WHEN THINGS GO WRONG Mistakes do happen. So a collaborative relationship with your translation company is essential for ironing out any problems. Incorrect translation If both translation and proofing are performed by the same organization, a reputable provider will use different translators to ensure independence. However, some clients may choose to use a different provider to proof the work, or even do the proofing in‐house by using someone from the regional office in the target market. Bear in mind that there is some permissible variation in translation as a result of personal preference. As long as the meaning is accurate and clear, and the tone is faithful to the original, a professional linguist will not make corrections purely on the basis of personal preference. Proofing by a native speaker who is not a professional translator may result in spurious corrections.
Page | 34
If errors are found in the translation, these should be forwarded to the original localization company for action. Bone fide errors are typically corrected free of charge. Culturally inappropriate material Idiomatic or culturally specific material such as jokes and cartoons should not simply be translated into the target language without considering how it will be received in the target country. The result may be misunderstood or worse, considered offensive. The localizer should identify such items during internationalization. Such risks are minimized by adopting the same sales and marketing strategies as you would in your own country; that is, the use of market surveys, focus groups and pilot releases. This paper is available at no cost from the Academy Translations web site. Please go to www.academyXL.com, enter your name and email address and download.
Page | 35