The work of 90s as I know it (sorry, no references, I was quite young then, do not care to loop up at the moment), was taking some corpus (and not an English on

The work of 90s as I know it (sorry, no references, I was quite young then, do not care to loop up at the moment), was taking some corpus (and not an English one) and letting a generative algorithm run on it. This was not a research but art, literature, and as such, actually could have been very regional. Of course the current state of art is on the different level but in basic principles not very different and as such not very surprising as an form of art. PS. it does not imply that no interesting results may grow out of this.