Standards

With all the talk lately about DITA (Darwin Information Typing Architecture), an open-source topic-based publishing standard, the use of XML for content development is finally taking off. DITA was adopted as an OASIS standard in May 2005. Large enterprises such as IBM, Nokia and Information Builders, Inc. (IBI) already use DITA, while popular document publishing applications such as Adobe FrameMaker, Idiom's Open Topic, and XMetaL Author, provide point-and-click interfaces that are helping to increase DITA's appeal.

XML is revolutionizing information management, yet terms within information have been largely ignored by the XML gurus. In fact, we have seen little more than the old HTML definition list model for back-of-the-book glossaries, and virtually nothing for terms in text, from even the most advanced XML document architectures. This lack of XML document markup for terms has prevented any serious repurposing of terminology for content management and localization.

On behalf of the LISA Terminology SIG, I have been lobbying to extend DITA so that terminology can be captured at the content authoring stage. For example, the existing definition list model is extremely limited: it only allows us to record a term and a definition. Today's needs are much more complex. We need descriptors such as part of speech and subject classification. These issues were fully explored in Developing Products for Multinational Markets: Effective Terminology Management, the workshop that I gave at the LISA Forum Europe in Zurich in November.

The DITA architects are listening. They recently submitted a proposal to the OASIS committee for a specification to handle terminology. The proposal addresses markup of terms embedded within sentences, as well as in separate glossaries. It also recommends extensions to handle more sophisticated terminology database entries, controlled vocabulary descriptions, concept associations, and lexical properties such as inflections for text analysis applications. (A fundamental requirement for the language industry is simply being able to indicate whether a term should be translated or not, which DITA fortunately already offers.)

This is great news for the language industry. When XML document formats such as DITA allow authors to mark up terms as soon as they type them, there will be nothing to prevent the more innovative content publishers from driving benefits through the localization process. Where there is a will, there will at last be a way.

---

Kara Warburton is responsible for IBM's terminology strategy, including tools, processes and data management. She is the chair of the LISA Terminology SIG, which defines best practices, and is a delegate to ISO TC 37, which defines terminology standards. She holds a Master's degree in Terminology from Université Laval and has held positions as translator and information developer. Warburton has published articles, given conference presentations and taught university courses on terminology. She can be reached at kara@ca.ibm.com.