Writing Engine - Part 1

Many years ago, slaving through a degree, a friend of mine endeavoured to build an artificial intelligence to create music. Surely a noble endeavour, I had yet to be sold on it being possible. It's all very well to toss concepts into the proverbial melting pot and "see what happens". But this is hardly the best way to get results. Like primitive alchemy, or some cultish belief that if they build it then it will do something, the AI musician never turned out a note that I heard as music.

It certainly turned out notes, but noise output is not music. Music has a structure, a logic, even an underlying beauty and symmetry.

It got me to thinking way back then about the early Apraphulians computer made of pulleys and knotted rope that unravelled to generate an automated play of cutouts, including scene changes and backdrop movement. Which led me to check the source.

I am appalled to admit this, but ... I have been duped by an April Fool (Apra-phul) joke. Oh, I feel so dirty in the scientific sense.

Regardless of the legitimacy of the history, the concept could be implemented with words!

I started then on my initial considerations back in about 1999 for what I am terming a Writing Engine. Some of the linguistic and computer science studies overlap into what is termed a Rational Engine.

One of my early writing engines coded in C++ used Generative Grammar and an artificial intelligence text mining approached to the network produced by words, phrases, sentences, and paragraphs. It could generate words in appropriate grammatical structures, and using windows of varying types, strings of sentences. Context and real meaning remained out of reach, unless hard-coded using an ancient information technology skill called "smoke and mirrors".

Semantic networks were still mostly theoretical and completely impractical at the time.

This concept would likely have gone completely untouched if I had not had a brain fart a few months ago. I don't know what triggered it, but, as some of my friends can testify, I had to smash out stuff on a white board that still now sits with scrawl all over it and numbered boundary curves with a legend on it.

I now have a structure for a 21st century version of what I had been intending to build. I am also teaching myself how to code mobile applications in Android. There are several good bodies of work that has taken place over the last 12 years, advancing key areas I needed to make the idea work.

So, this blog will assist as a sounding board, and perhaps provide some impetus to other similar projects that MUST be out there.