|
How do we translate the often chaotic, fleeting intersection of ideas in our brains into effective and efficient business systems? The answer lies in our ability to express (on the human side) and codify (on the machine side) complex concepts and models. Fundamentally, there are two approaches to bridging the human-thought/machine-instruction gap: - Translation – Translating between languages requires the ability to express meaningful statements in the source language, i.e. each statement must be evaluated. Based on this evaluation, equivalent statement(s) must be created in the target language. This requires the ability to correctly map syntax, vocabulary and semantics from the source language to a target language.
Translation is challenging for several reasons, including: (a) knowing two languages is a necessary but not sufficient condition for translation (domain knowledge and translation skills are also required), (b) languages have different degrees of ambiguity, (c) syntactically valid structures may have no meaning, (d) vocabulary terms often intersect and overlap, and (e) the boundary between vocabulary and semantics is fuzzy. A systematic approach to translation requires a meta-language to provide a foundation for translation. Future postings will discuss meta-languages developed by the IT industry.
- Common language – A common language avoids the need for translation. The discussion of Loglan in my last posting touches on some of the issues involved in creating a common language capable of supporting diverse constituencies.
A common language is challenging for several reasons, including: (a) most people want to use their favorite language vs. a common language, and (b) it’s difficult to design a single language that meets everyone’s requirements.
The Semantics of Business Vocabulary and Business Rules (SBVR) is an emerging standard that can be used as a meta-language to support translation, or as a common language to avoid the need for translation. By providing a mechanism for creating specifications in a natural language and leveraging formal logic, SBVR supports machine processing, thus bridging the gap between humans and machines.
By adding vocabulary and formal logic to traditional, syntax-based processing of digital content, SBVR and related efforts support fundamental improvements in how digital content is processed. There are two key advantages: solutions can now be expressed in natural language the resulting specifications are executable, which reduces or eliminates translation ambiguities and errors. I'll explore these topics in more detail in a future post. Meanwhile, if you have specific comments or questions about SBVR, Loglan, the semantic concepts discussed in this post, or related topics, feel free to contact me.
|
Data Governance