XML and DITA
The Key to Economical and Efficient Content Creation and Translation
Producing documentation for a global marketplace is a challenging proposition, especially for smaller companies. As most technical writers can attest, creating and delivering content in a multilingual environment increases demands on the documentation group. Often times, information is produced at the end of the product development cycle, under reduced time lines and tremendous pressure. Add to the mix numerous product iterations and multiple languages, and problems grow exponentially.
Consider this scenario, for example: if you are creating a separate user guide and service manual for 10 different product models, you will have a minimum of 20 deliverables. If each deliverable must be translated into just five languages, you now have 100 deliverables. And each time you require another language, you increase your deliverables by 20. This means that your documentation and localization departments are now responsible for creating and maintaining more than 100 different documents for just 10 products—and with a small staff.
Thus, document creation often impedes time to market for new products, as the documentation and translation lags behind. Moreover, with translation vendors charging upwards of 30 cents for each new word that is translated and 10 cents or more per word for exact matches, translation can be quite costly. Even minor wordsmith-type changes to documentation can wreak havoc on both productivity and the bottom line.
DEFINING AN INFORMATION MODEL
One of the key ingredients to resolving this problem is choosing the right information model. It is this model that determines how the organization controls, manages, translates, and delivers information.
Traditionally, businesses with multilingual requirements have managed the translation process in silos. A separate content set is maintained for each deliverable, which is then translated as a whole document independently of the others. This “whole document” approach, however, can mean additional translation costs. As mentioned above, minor changes or revisions can cause the entire document to be sent back through the translation process. It is easy to see how this whole document approach can quickly drive localization costs through the roof, even if you’re leveraging translation memory tools.
APPLYING THE DITA STANDARD
To avoid these challenges, and to handle such tasks in a more efficient manner, many technical publication teams are moving to the Darwin Information Type Architecture (DITA), an XML-based framework for authoring, managing, and distributing topic-oriented information. Using DITA allows documentation teams to modularize their content. Once the information is broken into modules (DITA topics), relationships (DITA maps) between the modules can be easily created and maintained, which can streamline the entire translation process.
By translating information in topics rather than whole documents, organizations can translate their content as each topic is completed. This greatly reduces translation and publishing delays, because information is being translated in a continuous stream instead of a mad rush at the end of the process.
Additionally, by maintaining proper relationships between the languages, no piece of content should ever be translated more than once. DITA-based modularized content creation and translation not only drives down translation costs, but helps the documentation and localization teams work together to produce materials faster.
Implementing the DITA model enables organizations to translate once, and then use those translations multiple times. For example, a large manufacturing company saved more than (US) $225,000.00 during a three-month period with one document set to be translated into 14 languages. The company previously had to produce multiple documentation sets and translate them individually as whole documents. Using DITA allowed the teams to write and translate information at a topic level, and then use those translated topics in various DITA maps to create customized outputs.
Once a topic is translated, it can be reused in multiple documents or deliverables. Since the translated topics can be reused in multiple DITA maps, there is no need to retranslate these topics every time they are included in new deliverables, as in the previous whole document approach. DITA maps are used to identify topics to include specific deliverables, and define hierarchical relationships between those topics. Additionally, metadata associated with these topics can be leveraged to create content for various personalized user needs. For example, you might have similar information in two deliverables, a user guide and a technical manual. The technical manual, however, is geared towards a slightly different audience. Instead of copying, pasting, and manipulating this information into each different iteration of the manual, the writer simply creates a different map for each project that includes the appropriate variations. This not only shortens the production cycle, it also reduces translation costs. And it ensures content consistency across different manuals.
Once this content is stored in an intelligent content management system (CMS), relationships can be maintained between all the topics in all of the languages. Following this workflow, any content identified as never having been translated would be sent to the appropriate translation tool and returned back to the CMS for publication, while the previously translated pieces would be used as is. As a result, far less content needs to be re-sent for translation.
Additionally, a DITA strategy combined with a content management workflow can greatly reduce click charges, since only new or updated topics are being run through the translation process. Say, for example, you are producing a user guide made up of 200 topics for a new DVD player. This DVD player is very similar to the rest of the product line, so only 75 of the topics are new or updated; the rest are reused from previously published documentation. Under the traditional “whole document” approach, the organization would have to send the entire user guide out to be translated, and would accumulate click charges for the 125 topics that had not changed. Using the outlined DITA strategy and tools, only the 75 new or changed topics are sent for translation, reducing the amount of content to be translated by 60 percent.
DEMONSTRATING THE BUSINESS CASE FOR DITA
The business case for moving to DITA is very strong, particularly for businesses with multiple products and selling into multiple countries. Creating a matrix of existing costs and potential savings areas should be fairly easy. Look for word count, per-word translation cost, and click charges. Next, begin to look across all content for reuse opportunities. This should reveal considerable possibilities for financial savings.
Consider this common business case scenario: an organization has created a manual comprised of 100 topics, with a total of 20,000 words. During the last production cycle, ten of those topics have changed, with five having minor rewrites and five requiring total rewrites. The existing translation cost is $.30 for new words, $.20 for fuzzy matches, and $.10 for exact matches. In an environment with no translation memory whatsoever, you would expect to pay for 20,000 words. Your cost to translate one document into just one language could easily be upwards of $6,000. Even if you apply translation memory tools, perhaps 25 percent of words being new, 25 percent fuzzy matches, and 50 percent exact matches, this would still cost approximately $2,500 to translate into one language.
If, instead, the company uses a solid DITA strategy, they will eliminate the click charges, so the cost would drop by at least $1,000. Assuming the updated and new topics use content references for linking content, the organization can further reduce its translatable content by approximately 30 percent. This means that the same document that cost $6,000 per language to translate will now cost less than $1,000 per language.
With ten documents in ten languages, the savings is $500,000 for just one document set. Since you can publish directly from a DITA map to print-ready PDF, HTML, and CHM file formats, these savings don’t take into consideration the fact that you’re virtually eliminating desktop publishing costs as well.
In addition to the bottom-line savings, using XML and DITA to create, translate, and manage content also has top-line impact. If documentation and localization teams are able to produce materials faster, this supports the goals of going to market in more languages and more regions faster—without additional resources!
TAKING THE NEXT STEPS IN YOUR ORGANIZATION
The first step is getting management to buy in. Using the savings calculation from the above example and assessing your own documentation is a start. Position these solutions in line with the company’s product and globalization goals.
Next, you need to engage other departments in the mission of more efficient content creation and translation. This means ensuring the documentation team and other content contributors and reviewers understand that writing in a topic-based environment saves time and money.
Using DITA to manage multilingual documentation makes good business sense. By moving from whole-document content creation and translation to topic-based writing and translation, organizations can significantly drive down the costs of translation, can ensure consistency of information worldwide, and can increase speed of content delivery.
ABOUT THE AUTHOR
Bret Freeman is a Principal Consultant for JustSystems’ XMetaL® Content Lifecycle Solutions, assisting organizations with their content strategies and technologies. Bret has more than 15 years of experience in content lifecycle operations with Fortune 500 companies, with expertise in the areas of enterprise content management, content migration, technical publishing, multilingual information management, collaboration, knowledge management, and complex workflow design.
ClientSide News Magazine - www.clientsidenews.com
Please see some ads as well as other content from TranslationDirectory.com: