ITC SIG Article: Managing Multiple-language Documents

Summarized from the ITC SIG list, December 2005

In October, 2005 Byron Hills of the Single Sourcing SIG asked if anyone has come across a reasonable method for managing, without using XML:

• Tracking of changes to an English source document
• Sending those changed sentences and paragraphs out for translation
• Getting the localized versions of the source document updated with the translated changes

Managing changes in multiple languages using FrameMaker

Carla Martinek responded, based on using FrameMaker on Windows. Carla's group uses specially formatted revision history markers in FrameMaker to indicate all changes to user guides. They generate a list of markers (LOM) file in FrameMaker, convert the results to a table, and then print the entire document to PDF. This provides a very useful reference for the translators and any reviewers, with hyperlinks to the exact spot that the changes occurred.

For those not familiar with FrameMaker, Carla explained that one uses markers to insert metadata and other information at specific locations. Carla's group created their own marker called Revision History, which is specifically for recording changes to content. Because the group shares files between multiple books to achieve a single source, it becomes very important to track where, when, and who made changes.

A LOM (List of Markers) is a generated file, similar to the Table of Contents or the Index. Once you have inserted markers into the document, generate a LOM file using those markers, and the result is something that resembles an index. When you print the entire book (including the LOM) to PDF, each entry is automatically hyperlinked. Place the LOM anywhere in the book; the back is a good location. The LOM must be printed into the same PDF file that the book is.

Carla comments that the hard part is to train yourself and others to put the markers in at the time the changes are made.

Carla further notes that changes could be tagged with conditional text that show the text as colored so that it can be easily found on the page. If the changes are small, such as a sentence, paragraph, or graphic here or there, the translation project manager extracts any content that changed, has it translated, reimports the changes into the last translated version of FrameMaker files, and generates new files. If the changes are anything more extensive than that, the translators run the entire document against the Translation Memory. While cost is a factor to consider, it is important that the translations are correct. Bad translations cost a lot more in the long run. For some languages, there may be 2 or 3 variations on a word, depending on how it is used or even the words used around it. Translation vendors may offer free or very reduced pricing on 100% matching phrases in documents that have been translated before and are simply being updated.

Managing multiple language documents within a global strategy

Brian Hertz (www.sdl.com) responded that all of the capabilities Byron asked about are part of a global information management strategy which requires four primary pieces:

• Authoring tools
• A file, content, and document management system
• A centralized approach to localization and translation
• A globalization management system

Tracking changes to an English source document

Tracking changes to a source document can be achieved in a number of ways. The simplest and least expensive method for tracking change would be to adopt a version control system using a central file server. There are several tools out in the market designed specifically for version control. These are mainly for managing documents in departments or work groups.

More robust ways to track changes cover the gamut of content management and document management solutions for the enterprise. CMS and DMS software from Documentum and Interwoven, among others, is powerful and expensive but well-suited to large corporations. The tools include sophisticated checkin and checkout and advanced XML capabilities.

Sending changed sentences and paragraphs out for translation

There is new technology on the market which can automate managing global content and the translation process. These new systems work with existing Translation Memory (TM) tools used by the localization companies by leveraging the previous translation and thereby reducing the cost of translation. These technologies also have tools to integrate with a CMS, DMS, database, or other document management system. In addition, there hooks into the authoring tools which can be used to maximize use of the TMs as the content is created.

Another key component of a GMS is a centralized repository for terminology. This is a critical component for consistency in use of multilingual terminology throughout a company.

Getting the localized versions of the source document updated with the translated changes

Updating and managing new translation can be achieved by use of the automated tools in the GMS. A GMS provides tools for project managers and reviewers and tracks their comments and changes throughout the process. The final documents can then be placed back into the CMS or DMS or can be stored on a central file server.

Most GMS solutions are native XML, but are capable of handling a variety of source file formats. However, many CMS and DMS systems do not require XML, and a fileserver solution does not require the use of XML.