Monday, October 13, 2008
A couple of thoughts
I just wanted to weigh in a bit on some of the recent discussion about Issue 1 and maybe help to make a couple of things clear.
First off, Issue 1 is not a hoax nor a prank. It is (and I think should have been immediately seen as) parody, parody as obvious as anything on The Onion. If there is anything amusing about the project, it is that so many folks did not see that. Maybe that's because it is bad parody. I'm too close to the project to be objective--I'll leave it for others to make that judgement. But speaking for myself, there was no intent to fool anybody, just to evoke a chuckle or two.
Second, the ETC project is, and should be discussed as, separate from Issue 1. I was multiply motivated to continue the project after completing my thesis. I did not feel as if the thesis satisfactorily answered the question as to whether machine poetry could compete with traditional poetry. I had attempted to devise and implement some controlled testing, but could not for the life of me devise an
adequate null hypothesis, let alone an experiment that would reject it. So after some considerable time, I decided that the only way to test was to actually send out the work and see what happened. It was important that in sending poems out I not identify them as machine works because that would irreparably compromise the experiment. Some editors would accept the work only because it was borne of the machine and others would reject it for the same reason. So Erica was born. And of course that kind of exercise does have at least some of the characteristics of the hoax. And I confess to some pleasure in the act.
And as are most alternative artists, I wanted to be disruptive. That motivation, at least, has been amply satisfied in the last two weeks. (I know: I contradict myself. I do that a lot.)
But there is another, to me, more important motivation, which speaks squarely to Issue 1. And that has to do with the broader community of computational artists, particularly those working with text. A problem confronting these artists is where to get text to support their work, especially since the demands of an artifact capable of processing thousands of elements per second and storing gigabytes of data require enormous amounts of it. It is physically impossible to manually write the 1000s of pages needed to support certain types of work. Further, developing excellence in the skill sets required for developing computational artifacts and literary artifacts would require at least double the effort it takes to become either a
highly-skilled technician or highly-skilled author. My thoughts were that artificially generated texts could be used in such works. (One of the reasons I wrote the most recent version in Java was to facilitate such usage--and also why I've posted the source.) Erica never gets tired, never complains and works for just about nothing. So far only the Issue 1 guys have taken me up on that.
Finally, all of this is past. I have turned to other interests, none of them computational, and at this time, have no ambitions toward furthering the project (another reason for releasing the source). If the project has value, someone else will pick it up. If not, no one will. BTW: There is a clear line along which Erica's poetry can be improved significantly, which does not require any programming knowledge whatsoever. Just a little Xml.
If anyone is interested, I am happy to respond to questions about the software's design. Just email me.