Unicode StandardConference BoardConference CDLast ConferencePast ConferencesNext Conference

Improving Initial Translation Quality: Correlation of Context

David Kumhyr - IBM Corporation

Intended Audience: Content Developers, Managers, Software Engineers, Technical Writers, Translators, Project Managers
Session Level: Intermediate, Advanced

Intended audience: Intermediate with knowledge of the software development cycle and familiarity with translation and verification of software.

In this presentation I describe a method of improving the initial translation quality that will accelerate the product schedule. Using this method can improve the initial translation quality so that a translation verification testing would not be necessary prior to product release. This saves a large portion of the cost of software product translation and removes weeks from the translation and test schedule. It is a radical approach, though one that we have begun to implement.

If one were translating a novel the context for creating an accurate translation is derived from the work and the translator's understanding of the novel's setting; the historical, cultural, location and the socioeconomic strata of the novel. Context is the all important aspect for understanding the work to be translated, and the basis from which an accurate translation arises. The translation of software products is more difficult than a novel; compounded by a number of factors. The setting of the application is a creation of the interface designer the translator has little experience in this 'setting' to relate to and that context is derived from. The translator may understand software and be conversant with general computing but lack specialized knowledge of the product area such as systems or IT management. Translation difficulty is also compounded by the developers' use of good internationalization practice. By moving the human language out of the program interface and into resource files the text that is to be translated is disassociated from the interface. This leaves the translator to guess which text string in a file will be associated with which interface element.

To solve this problem and to create a good initial translation some context must be provided to the translator. Most methods for presenting context to a remote translator present major difficulties. However, if we were able to provide accurate and understandable contextual reference for the application text and provide a means of relating text in resource files to the program interface the initial translation would be of higher accuracy. With an initial translation of higher quality the remaining defects in translation would likely be spelling and grammar related. Use of automated spelling and grammar checkers followed by an editorial review would be sufficient remedy to enable the production of a high quality translation. The translated output from this process could be of high enough quality for integration into the product for general release without further testing.

When the world wants to talk, it speaks Unicode

Unicode StandardConference BoardConference CDLast ConferencePast ConferencesNext Conference
International Unicode Conferences are organized by Global Meeting Services, Inc., (GMS). GMS is pleased to be able to offer the International Unicode Conferences under an exclusive license granted by the Unicode Consortium. All responsibility for conference finances and operations is borne by GMS. The independent conference board serves solely at the pleasure of GMS and is composed of volunteers active in Unicode and in international software development. All inquiries regarding International Unicode Conferences should be addressed to info@global-conference.com.

Unicode and the Unicode logo are registered trademarks of Unicode, Inc. Used with permission.

12 December 2002, Webmaster