it's no secret that parallax has been withering on the vine for quiet some time. originally, i had intended the site for "guest" bloggers. exactly one of you might remember way back in the day when kristine posted on environmental and sustainability issues. but unfortunately she didn't have the stamina for daily blogging and eventually gave up. after collecting dust for a while, matt started posting on design and usability topics, but that didn't last long either.

so, i've finally taken back control of parallax and decided to try and experiment that i've been wanting to do for a long time. the parallax posts are now automatically generated from text harvested from "the vast wasteland" . as of today, the dataset is the entire archives for ex_machina which go back almost 4 years. the initial results are promising and somewhat amusing with almost no tuning of the scripts or massaging of the generated content.

while i've wanted to do something like this for a long time, it was after seeing the markov blogger code that i decided i should start playing around with the idea. i decided to experiment with colin frayn's excellent random sentence generator because, well, i'm lazy and it required less up-front work for me to get running. after doing more tuning, i think the next task will be to figure out how to get it keep links, like the markov blogger. and then i'll fold in support for automated posting to blogger, so i can run it in the background at random times through the day.

why do this? well, i've got a lot of content sitting around and i've been wanting to do something with it that's more fun than the "on this day 3 years ago" type of stuff. and isn't everyone interested in computational linguistics ?