Wednesday, February 13, 2013

The World Outlook for Wood Toilet Seats

by Drew Martin
There is an interesting "Techwise Conversation" on IEEE Spectrum, Is Micropublishing the Death of Publishing – or Its Salvation? It is conducted by Steven Cherry with Thad McIlroy, James Morrison, and Philip Parker. McIlroy is an analyst and consultant for the electronic publishing industry. Morrison is a writer/editor/designer, and the brilliant Australian behind The Caustic Cover Critic blog. Parker is a professor of marketing at INSEAD, and author of hundreds of thousands of titles by way of his patent (7266727), which supports his invention for,

“Automatic authoring, marketing, and/or distributing of title material. A computer automatically authors the material."

 At first (the first few times I listened to his part of the interview) I could not wrap my head around what he was talking about. It seemed so post-human. Parker creates algorithms that create books from cover to cover in a matter of minutes, including a table of contents and supporting charts for the relevant information. The algorithms mimic the process of how an economist would gather and compile data. Cherry questions Parker whether his books with titles such as Webster’s Slovak-English Dictionary and 2007-2012 World Outlook for Wood Toilet Seats (which sounds totally absurd) are typical. Parker replies,

"Yeah, actually they are pretty typical. The publications are going after the long tail, or niche, topics that the publishing industry would normally not cover. And so for the toilet seat cover one, the inspiration for those reports came from microenterprises, mostly based on projects I was doing in Haiti and other countries, where small enterprises literally make very, very specific items. So if they want high-end market research or anything analytic, they would have to literally be down to that specific level. People don’t make products; they make very specific products. So the reports cover minutiae topics because that’s what importers and exporters generally sell."

Parker actually makes more and more sense the deeper you look into his projects. In Africa, he created 24/7 automated weather reports for people who had never heard a broadcast of the weather in their native language.

The most interesting part of the interview was Parker's insight on genre,

"Yeah. The basic idea, this is kind of the thought process of automation, is that you don’t create a software to write books or videos or PC games. We do all formats, by the way; it’s not just books. But you do it by genre, so you have to say, well, fiction versus nonfiction, and you could say, well, let’s do nonfiction. And then within nonfiction, well, there’s different types. There’s genres, there’s bibliography, there’s biographies, there’s crossword puzzles, there’s dictionaries. And then what you have to do is you say, no, that’s not really true. There’s no such thing as a bibliography; there’s an annotated bibliography. And then you say, no, there’s not really an annotated bibliography; there’s a sub-subgenre called Cambridge-formatted annotated bibliography, etc., etc.

And as you drill down within a genre and you can no longer find a subgenre, typically at that level you’ll discover that authors, the people who do them manually, actually follow very formulaic patterns. And that’s true for poetry and other types of writing styles as well. And so that’s what I meant by “genre,” is that you drill down to the point at which there’s a formulaic approach to authoring in that domain."

This leads to a conversation about poetry, with Parker commenting,

"Right. Within literature, of course, there’s subgenres of literature, poetry being one of them. And then you say, well, there’s no such thing as poetry; there are subgenres of poetry. And you could say sonnet, and you can say, no, no, no, there’s subgenres of sonnets. There’s metasonnets, there’s—etc. And when you get down to the very finite level, even within poetry, you have highly formulaic approaches. Like as in a sonnet we know that it has a rhyming pattern of AB, AB, CD, CD, EF, EF, and then GG in the final couplet. The first line is a question. The ninth line is a turn, which usually has the word but or yet or since, something like that. They’re somatic, and we decided to do metapoetry, where the poem is about the poem itself or about the author of the poem.

Shakespeare wrote a sonnet, No. 76, which was a type of metapoem, where he kind of complained a little bit that he was stuck using the same formula in writing his own poems. So it’s a metapoem because it’s about himself and the poem itself. We took that concept and created algorithms of a computer program writing sonnets, stating it cannot write a sonnet properly, but doing so, it writes a sonnet properly. So it follows the iambic pentameter formulation, has the exact correct number of lines. It’s highly constrained writing once you get down to that level."