by Jana Bily, STC Boston Chapter International SIG Leader
According to Byte Level Research, by 2010, 79% of all Internet users will be non-English speakers. Many companies have already recognized this fact and offer their websites in multiple languages. For others, website localization will soon be an important and necessary task to undertake. Successful website localization depends on a well-fitting site architecture, an optimal content management system (CMS), and an efficient localization workflow.
Content Management Systems
Website localization starts with designing an efficient web architecture that can easily hold and manage the website’s content in multiple languages. In recent years, CMSs have gained in popularity as organized repositories of information. CMSs range from very simple to very complex data storage media and their infrastructures allow for the creation, modification, archiving and removal of information resources. Not all of them, however, are localization-ready and support foreign languages.
Selecting the right CMS from the hundreds that are available on the market can be a daunting task. For example, CMS Matrix (www.cmsmatrix.org), a community service to everyone interested in looking for a means to manage website content and a member of the Compare Stuff Network, lists 791 of them!
For small organizations, it is not practical or affordable to deploy a large CMS such as Vignette or Documentum, so they rely on simple, open-source systems. Luckily there is a plethora of them available — Joomla!, Drupal, BitWeawer, XOOPS, and Typo3 to name just a very few.
Some of the open-source CMSs have built-in functionality to manage multilingual content. For professional multi-lingual websites, this is a minimum requirement. If your website content is relatively simple in terms of formatting (i.e., it contains embedded tags) and linguistics, will remain quite stable, and you only plan for limited content updates, then an “out-of-the-box” CMS is for you. However, such a system will have only limited functionality whereas they provide a web-based editor allowing you to translate your content online. Soon you may find that this is insufficient.
If your source content format is fairly rich, the subject linguistically elaborate, or the updates frequent, you may need a more robust multilingual translation solution. Your CMS will store the multilingual content but you will need a more elaborate environment and methodology to create and maintain the translated content.
A good solution is to use the CMS in conjunction with the translation memory and automated connectors.
Translation memory is a computer-aided translation (CAT) technology that allows for an efficient and consistent reuse of translated content. It has been widely adopted by the localization industry because it effectively decreases the translation cost and yields consistent translation quality.
CAT tools are powerful optimized translation environments for the translators. The tools work in conjunction with electronic dictionaries and store translated content in databases for reuse. When translating content updates, the tools will recuperate what has been translated previously, thus only the delta is left to be translated by the translator. While a CMS typically just tells you that a source article has been updated, a CAT tool will allow you to update only that content that is new or has changed.
If your website is large and contains rich content that changes frequently, you will soon find a need for a connector. A connector is a piece of code that automates the exchange of content between the content repository and the translation environment. It is typically designed to facilitate the following tasks:
- It extracts the content in a structured and systematic manner (using generated IDs) while it preserves the source formatting.
- It pushes the content through the translation process.
- It reintegrates the translated content and makes it ready for publishing.
A Final Note on the Translation Quality
CMS and translation tools have been designed to facilitate the management and translation of websites’ content. However, much of the work still depends on humans. Producing high-quality translations is teamwork of translators, reviewers, subject matter experts, and localization specialists.
About the Author
Jana Bily is the Director of Sales and Marketing at EzGlobe (www.ezglobe.com), a full service localization company. She draws on her experience from having managed complex multi-language, multi-million-word localization projects at PictureTel, Rational Software and IBM. She can be reached at email@example.com.
This is a reprint of an article originally made available in the September 2007 issue of Broadside (STC Boston newsletter).