Who

My name is Will Crowther (no, not that one) and I became interested in the natural language / computer interface many years ago when a friend participated in the Loebner Prize. I was just had recently transitioned to a role as a full-stack .NET C# Programmer and found the topic fascinating and something to pursue as I got a better understanding of more advanced programming. My first attempt was working on an XML-based Natural Language solution that was a great way for me to tinker and explore with with C# and SQL. It accomplished some simple tasks but was not yet anything ground-breaking.

About this time I got interested in other projects, including one I created called SeedPacket.net that leveraged a custom Rules Engine to generate random data using Linq syntax. This was about the time that Siri and her friends were starting to become popular. My assumption had been that this new voice-interface technology would take off and become a much more central part of our computing environment. Ten years on, reality was a bit more complicated than what we might have expected. The voice interfaces could only handle simple commands but have not yet developed personalities. Much of the technology was still not accessable to the average programmer. Bots are not that interesting to talk to. Much of the fun is seeing how little they seem understand and what tricks that they have been programmed with to make you not notice. These days, computer games have incredible environments but do not yet have NPCs that you can talk with.

With these revelations, I recently got interested in developing MachineEnglish again. I realized that creating a powerful new Natural Language parser is a tall order. To accomplish that, you need to be able to harness a broad range of insight into the human mind and how it processes language. So create a tool that allows everybody to contribute and embed their insights. This means the technology will need to be easy to understand and transparent to people with varying technology backgrounds. In order for it to develope and evolve quickly it needs to be very adaptable, something between a programming language and a dictionary (or encyclopedia?). It needs to be very testable so that you can define how it works, but eventually be able to evolve enough to escape programming tricks to become truly interesting and even artistic.

If you find Machine English interesting please drop me a line at wcrowther22@gmail.com.