Language Workbenches

Wouldn't ya know it! On the very same day that I blogged about Customer-Oriented Requirements Architecture (CORA) as The Next Big Thing, it turns out Martin Fowler wrote an article about Language Workbenches which seems to be getting at exactly the same core idea:

• Language Workbenches utilize things like Meta-Programming Systems and Domain-Specific Languages (DSLs) to let the developer work more closely in the conceptual domain of the various subject-matter "spaces" of the requirements and the design.


• They provide all sorts of IDE and refactoring support for the language domain they are created to support.


• It seems a bit more focused on the design-end, whereas my CORA idea is a bit more focused on applying architectural principles and design patterns to the expression and maintenance of the requirements. I believe the end-result is the same however.

The more enabled we become at formally expressing the requirements in a language and framework more closely bound to the problem domain, the more important it will be to apply principles, patterns, and practices of refactoring, encapsulation, modularity, etc. to the groupings of requirements we develop and the relationships within and between them. And Language Workbenches become part of the environment that supports, maintains, and automates requirements dependency management and traceability

Feature-Driven Development (FDD) has an interesting way of trying to do some of this with its feature-sets and color modeling patterns. See recent discussions on the agilemanagement YahooGroup and the newly created colormodeling YahooGroup for more details.

I may blog in the future about the relationship between FDD, Color-modeling, "grammar rules" for domain-modeling, DSLs, and the Law of Demeter.