When I was a kid, I loved building things. I made boats, and model rockets, and a photo darkroom with a homemade enlarger. I taught myself to program after I graduated college with a degree in Physics. I wanted to create natural language processing software, so I taught myself LISP and Prolog, the two dominant AI languages in the 1980s.

It turns out that symbolic processing of natural languages is hard to do, and expert systems didn't really work either. So, the industry entered the so called AI winter by the end of the 80s.

I repurposed my parsing skills to working on programming languages. I got a job creating SQL parsers for a software company in Silicon Valley and moved my family to Mountain View. That was 1990. I've lived here in Mountain View ever since.

For two decades I mostly worked on Javascript related languages. I served on the Javascript standard committee for ten years from 1998 to 2008 during the height of the browser wars that largely revolved around the design of Javascript.

In 2011, after an 8 year stint at Macromedia/Adobe as a lead on the design and implementation of Actionscript 3, I started Artcompiler. Artcompiler was started on the premise that it is better to create software by allowing experts in the problem domain to engage deeply in the code. And that this would be possible if there was an application specific language at the core of the application.

Starting in about 2008, I started experimenting with this idea in an open source project called Graffiticode. The idea was to create web APIs that take an argument that is code in a DSL. This code defines the behavior of the API.

Fast forward to now. The applications are full stack micro-SaaS apps. Every app has a task specific language used to program it so that domain experts can participate deeply in the software creation process. Artcompiler is how we deliver on the promise of this vision.