About reproducibility, paper or screen is somewhat irrelevant in the perspective I was presenting, since the bottom line is processing; that component is what prevents certain works from being archived with current approaches… I know, I know, there are examples and counter examples, so this discussion has to be put in context.

About IA – AI, I meant IA, since Literatronica was developed under the support of CAVIIAR (Centro Avanzado de Investigación en Inteligencia Artificial); IA is the Spanish acronym for AI… I know, I know, this is a post in English; nevertheless I allow myself the nostalgia of my mother tongue. Additionally, the model and applications of Literatronica have been shown to Spanish audiences since 2002 (and its precedents since 1996). I know these news did not make it to the English speaking community until recently.

About the metric, yes there is a metric, yes the triangle inequality holds, no the author does not need to define a N*N network of links; about the latter, there is precisely where the IA system plays a role: building links on the fly that guarantee narrative continuity, thus reducing the topological effort from the author.

About the HT08 paper, well, I already submitted a related paper about mathematical modeling of narrative spaces, which is related to Literatronica; we’ll see if it is accepted ;) Literatronica is not a “how to do it,” but rather a “how I did it.” I believe it is important for the diversity of the field. I will show the guts of Literatronica and an application (the GPS – Global Poetic System for the City of Barcelona, with Laura Borràs) at ELO’s Visionary Landscapes in Vancouver next May.

Hope to see you in Vancouver… and hope to see you in Pittsburgh. I am looking forward for the opportunity of discussing all these matters in person.

But the hypertext writer added some links, and omitted others that she might have added. Just as a bard might extend some scenes and foreshorten others, the hypertext writer has (we hope) carefully arranged the links and anticipated your reading.

You write that “I don’t want to read about the break-up while they’re still dating.” That might be true, but we have good reasons — we’ve always had good reasons — for telling stories out of the sequence in which the occur. Homer urges the muse to sing of the wrath of Achilles, but he isn’t angry yet. The first line of Love Story asks “What can you say about a girl that died?” But she hasn’t died yet, they haven’t even begun to date yet.

Hypertext writers control pacing, plot and structure with language, and with links.

Now that I understand! As a tentative reader (with the mind of a writer), I have been sifting through some typertext works. I worry constantly that I’m not getting the essence of the story because I don’t know how the writer crafted it. With a paper book you assume you are reading in the correct order, the way the writer wants you to learn the story. But Hypertext? What if I’m skipping ahead? I don’t want to read about the break-up while they’re still dating! How to you control pacing, plot, and structure? Or is this medium only suitable for certain stories?

I agree.

I’d also suggest that the reproduce-ability issue goes to what I would call Nelson’s First Law of Hypertext: it’s computer mediated and cannot be anything else and still be itself.

Reproduction in paper is a paper tiger.

The sophistication of the Literatronic IA (or do you mean AI) engine is, I think, beside the point; we’re talking about two related formalisms whose consequences are, at this point, unclear. (Question: do the distances in Literatonic induce a metric in mathematical terms? Does the triangle inequality hold? And, if a work has N writing spaces, does the author need to specify N*N distances?)

It would be great to see more about Literatronic in a HT08 paper!

I think this post will provide some context.