DSLs Lead Development Paradigm Shift

Domain specific languages are leading the way to new programming models.

LAS VEGAS-The software development world needs to move from the current phase of using static, procedural languages and frameworks to one of language-oriented programming.

Speaking at TheServerSide Java Symposium here on March 26, Neal Ford, senior application architect at ThoughtWorks, discussed the concept of DSL (domain-specific languages) and how they are changing the development landscape.

A DSL is a programming language designed for-and intended to be useful for-a specific kind of task. DSLs focus on doing one kind of task well.

Indeed, Ford cited one of his ThoughtWorks colleagues, Ola Bini, a Ruby enthusiast, as describing his view of a future stack for developers as one consisting of a "stable language" as the bottom tier, dynamic languages on top of that, and DSLs as the top layer of tools in the futuristic programmer's basic toolbox.

Ford also promoted the concept of "polyglot programming" or the process of leveraging existing platforms with languages suited to particular problems.

ThoughtWorks' tide is rising with a new product. Click here to read more.

"It is time for enlightenment to help us through the frameworks," Ford said, adding that DSLs improve the software development process by "eliminating noise." Ford was referring to application development frameworks such as Ruby on Rails.

Ford said programmers experienced in dynamic languages tend to build DSLs on top of their low-level language. "Using DSLs evolves the way we build and use frameworks, escalating our abstraction levels closer to the problem domains and farther from implementation details," he said.

"It's evolutionary, not revolutionary; DSLs are a much better abstraction than XML," Ford said.

Ford referred to external DSLs as "roll your own language," which used to be done with "old school tools like Lex and Yacc"-two Unix utilities for programmers-that were "pretty primitive," he said.

Newer tools like ANTLRWorks are better, but still not great, he said. ANTLRWorks is based on ANTLR (Another Tool for Language Recognition).

"But we're getting to things like -language workbenches,'" Ford said. He said companies like Intentional Software and Microsoft, with its Software Factories technology, have early implementations of language workbenches, but perhaps the best example to date is JetBrains' MPS.

MPS is JetBrains' Meta Programming System, which "is a new programming environment which makes it easy to define your own specialized languages, and use them together with any other language, as you see fit," according to a description on the JetBrains site.

"Your languages also have full IDE [integrated development environment] support: code completion, navigation, refactoring, and more, and you can add your own specialized support (such as special editors) if you wish," the MPS description said. "You can even extend existing languages and add your own features. MPS eliminates the programmer's dependence on languages and environments, giving more freedom and power to the programmer, ultimately making programming easier, more fun, and more productive," according to the site.

Should developers have morality in their coding? Click here to read more.

Moreover, MPS is an implementation of Language Oriented Programming, whose goal is to make defining languages as natural and easy as defining classes and methods is today, JetBrains officials said.

"They're going to open-source this later this year," Ford said of JetBrains and MPS. "This is the first wave of these language workbenches that are going to come out."

Ford said Java developers adopted XML for data binding and began to use X M L as an external DSL because it is easy to parse, "but the question comes up: how much is too much," he said.

Ford added that raw X M L code can be frightening. "It frightens non-developers; it looks worse than Java code," he said.

However, "XML is marginally OK as a data format," he said.