User friendly business rules run on a scheduled basis to find new or changed content and once found, create translation projects. Those projects use the WorldServer workflow engine to shepherd the content through the globalization life cycle. Workflows are comprised of human steps and automatic actions. Human step assignments are designated to workflow roles or individuals in the workflow or at the time of project creation. At any point, WorldServer can send automated notifications to step owners or other interested parties.

Assets are first segmented into translatable and non-translatable segments using filters created for specific file types. WorldServer supports a variety of content types: database records, software property files, HTML, practically any form of XML, printed file types such as InDesign INX or FrameMaker MIF, and of course, Microsoft Office documents. Each filter determines which elements are translatable, non-translatable, or embedded within other translatable elements.

After initial segmentation, the system checks the existing translation memory to see if any of the translatable segments already exist for the given target language. Perfect, or 100%, matches are automatically leveraged and applied to the translatable content. Of these perfect matches, some are also In-Context Exact, or ICE, matches. These matches appear in the same contextual order in the translation memory as they do in the translatable content and therefore do not need to be reviewed by a translator. Over time, as translation memory grows at the completion of every project, more and more matches will be found during the leverage process, driving down the overall cost of translation.

WorldServer solves the problem of including not just changed content for a translator, but also the surrounding content to give the translator context. With a mature translation memory, when you send content out for translation, you don't have to worry about hiding content from the translator, since the content that has not changed will be automatically translated, leaving only the changed content—in context—to be translated. The translators can look at as much context as they want without you having to worry about paying more.

WorldServer integrates terminology and context viewing in the entire translation process. This ensures that key concepts, brands, ideas, etc. for your organization are preserved throughout and that translators can see approximately how their translations will appear in the final output.

During the leveraging process, even if a perfect match is found, it should be reviewed and approved by the content owners to ensure quality. Legacy procedures outside of WorldServer might do this during project archiving by creating golden translation memories (TMs). WorldServer provides segment review status for each translatable segment and their counterpart stored in the translation memory. This way no archiving or specific golden TM maintenance is required. In effect, by reviewing content during the natural project workflow, golden TM is built automatically and contained in the same TM used for the project.

WorldServer provides many out-of-the-box connectors to existing content repositories, such as SQL databases and Perforce. On the backend WorldServer integrates with MySQL and Oracle for database storage of all translation memory and WorldServer configuration data. Tomcat (included with the WorldServer installation package), WebSphere, and WebLogic are available as support web application servers. Even with such a wide array of supported platform configurations, SDL also exposes core APIs via the Software Development Kit that can be used to create custom integrations or SOAP standard Web Services connectors.