Computer system builders increasingly draw on a wide variety of programming languages to assemble systems. From database queries in SQL, to core, CPU-intensive codes in CUDA, to GUIs in JavaScript, and with python to wire it all together, these builders routinely benefit from a close match between the language's form of expression and the wide variety of tasks they perform. And yet, the current status quo merely scratches the surface of what is possible. If systems were designed by first thinking about the best possible language to design (some part) of their system, building that language, and then using that language (and others) to build the system, these builders could realize tremendous productivity gains.

 

In this talk, I will try to bring across how language-oriented programming holds the promise of making programmers more effective and making more people programmers. I will illustrate the idea with a few example programming languages, and then try to give a sense of how the programming language Racket provides a particularly useful set of technology for building systems using the language-oriented programming philosophy.