Markovian Parallax Generate: On digital writing and poetics

Check out the Random Blog Finds

Posted in Source Material (input) by Eric Goddard-Scovel on August 29, 2007

I’m still easing (ha! anything but, really) into this third semester of grad school.  Until I finally start updating this site again, go take a look at some of the blogs I’ve linked far down on the right sidebar under Random Blog Finds.  These blogs often end up being used as source material for mchain and gnoetry.  Something about each one of them drew me in (I’m especially keen to tone and quirkyness).  The whole blog phenomenon is still exciting to me, and I love finding new oddities with the Random Blogger Page function.

Tell me what you find interesting/moving/amusing about any of them.

Source Material: What it is, where it comes from, how I find it, and how I intend to use it

Posted in Source Material (input) by Eric Goddard-Scovel on June 28, 2007

Okay, so It’s been slow going with the whole Markov chain writing process, but things are getting going. The input file I used with my last posted poem draft a few days ago has been updated a bit. I have corrected it for odd copy-and-paste errors and cataloged all of the links to the original source websites. [Note to self: do this while you copy the web sites so you don’t have to spend hours searching Google for links in 180 page document again.]

First thing, here’s the link to the input file: Mchain Input. I had to save it as a .doc file to upload it here, although it has to be in plain text format to work with mchain program.

What It Is and Where It Comes From

It is a 179 page collection of blog posts, livejournal entries, web pages, dream journals and dream interpretations. Some parts of the file are my own writing, mostly memories or dreams that came to mind as I was reading and adding texts to the input file.

And now, where it came from. Yeah. Here is the list of bookmarks. It is complete except for two entries which had either been removed or made private since I copied them:

If any of the authors of these blogs or websites are wondering why I’ve linked to them, I hope the rest of this post explains it. The subtitle of the blog might help with that a little bit too: On the use of Markov chains in the composition of Poetry.

How I Intend to Use It

Essentially, I am using the text of all of these websites to create a large database for the purpose of generating semi-random nonsense with a computer program that uses the Markov chaining algorithm. I use this nonsense to give me material for poems, which are hopefully influenced by the ideas and emotions of the original texts as well as the mangled nonsense text. It’s all a bit unclear at the moment in terms of rights to the text. Does the nonsense count as a derivative work, and with all of the texts becoming separated word by word (or chain by chain) and recombined, of whose work is it derivative? Does it resemble the original enough to any longer matter? I don’t know. It becomes difficult to tell.

So, if any of the authors of the original texts are uncomfortable with their writings being used in this way, please let me know and I will remove them from the file and the list of links above.

How I Find the input texts

I use Google Blog Search, entering a search query of several words like so:

+mountain +tarp +feel

I usually only use three terms, but sometimes I use more. I can’t remember what words I chose for many of these posts. I did focus on finding dreams, though, so I did a search that was something like:

“dream last night” or +”last night” +dream, or +”last night” +(dream|dreamed)

Seriously, if you’re unfamiliar with boolean, etc. search syntax, you should check the links below out. They cover it pretty decently, though I can’t yet find good info on the “|” (which means OR), or on filetype: searching. Good search technique makes things a lot more interesting.

Internet Tutorials: Boolean Searching on the Internet

About.com: Top ten web search tricks

About.com: How to find and search the invisible web

I also use the Random Blog functions on WordPress and Blogger to find odd pages. It’s tedious, but works pretty well sometimes.

There are other ways to find good blog sources, as well as other texts to be used as input. I could use hundreds of pages of my own freewriting, if I had written such a body of work. There are open license databases online, and one could put all kinds of novels and other books into the program. I prefer the down-to-earth voices of many of these blogs, as well as the spice of odd sites added in here and there. The quality of writing going on right now in the blogosphere is a mixed bag, but there is a lot of great stuff out there; highly personal at times, sometimes creative, and occasionally very moving.