this.predicate // metaforms

So I’m running into some problems with the code for my project, mostly due to my own programming inexperience. I’ve overhauled the basic premise so that it’s turning out a fairly smooth, real-time, 3D form based off of typed input. And I’m sure I can comb through the text and cue different actions based off of basic grammatical cues. But I’ve been unable to add the actions that would provide a meaningful representation of the written piece. No curves, no delay-tracking, no character generation, no line weight variation.

Taking the code back to basics this weekend, again, with the hopes of introducing ArrayLists for more dynamic management of objects. That will allow me to target specific segments of the object more easily, and hopefully, add some depth to ways that I parse the text information. Also working on the actual placement method for the segments representing words and letters. The hope is that I can drop in a few different objects, in place of the lines, that can support a greater graphical variance in width, height, and length.

I’ve been knocking around a few ideas for the story that will drive my metaforms project, but so far nothings really taken shape, so to speak. I’m still focused on Invisible Cities for a few reasons. It’s about the right length for the program I’ve set up. And more importantly, it’s a good balance between the abstract philosophical that lifts a story and the detail focused style that brings it within reach.

For my piece, I’ll be focusing a little more on NYC than a series of mythical cities. And the themes that arise in each city will become my personal interactions with the city. To begin with, I’m trying to go through a short series of reflections about my life here, both rounds, and peg down the little pseudo-zeitgeists. Each is about the shape of my life, and where possible, its relation to language.

First thing: need another name for the project.

Otherwise, the feedback from the midterm went well in both ICM and Metaforms, although it didn’t give me too much more in regards to how I can continue to build on the project. What’s clear is that I need to continue researching which (con)textual markers resonate the most and do a better job of communicating them visually. A friend referred to this project as a narrative fingerprint, which I like more than the original rationale of cityscape shapes growing out of stories.

But I guess the key question is how to involve people in the project more. At this point, it’s far too flat — a simple notation of what is, was, or will be. I like an aspect of that, but it doesn’t capture the level of ethos I’d like. So it looks like I will have to take the interactive route.

The most immediate way to do this is to allow real-time input of text and watch the shape build dynamically as typed. That also adds a layer of data to be tracked, including input tempo. That hooks in nicely to Nicholas’ idea of creating an app that textualizes and weights the conversational patterns in people’s writing/speaking. It also lends itself to the fingerprint idea. But I don’t think it really captures the ethos still.

I’m beginning to wonder if this needs to go two different directions for ICM and Metaforms. The latter should become a 3d shape — if remotely possible — via the Rapid Prototyping Service at the Advanced Media Studio. I’m thinking a relatively small object, hopefully multiple ones. This is definitely an external object to be examined — I’m not building some crazy architectural structure. I’ll have to give a little thought to how I’ll connect the text and shapes once it’s removed from the monitor. Can’t rely on interactive rollovers or anything, but I’ll still need to point out key transitions; sentences, graphs, etc. The former might be something a little more interactive. I’d like to set up a real-time speech to text conversion that visualizes speech. That might shift how I want to visualize this (an audio input will naturally take a different shape) but it could be fun nonetheless.

So next steps:

1. Translate current code into class/object-based
2. Add 3D/OpenGL code
3. Test speech-text library 

I need to identify the different types of data, both in the input (narrative/writing) and the output (visualization), and assign some sort of logic to the mapping. The challenge will be to get some sort of emotive quality out of the writing, but I’d like to avoid marking it up with my own interpretation. At the end of the day, this should be a tool that translates the hidden form of the narrative language into visual language. Or more succinctly, it will translate semantics to semiotics. In a Derridian/postmodern sense, this is meant more to open a door than arrive at a conclusion. It’s the opening salvo to a multimedia reinterpretation of narrative. (On a side note, when I begin looking at narrative interpretation, Heidegger’s concept of Destruktion might be worth pursuing.) 

Visually, this will be linear-based to represent the nature of a story and its relationship to the written word. (For now, I’m thinking plain lines will work, but I may consider morphing these into lines built or constructed out of shapes/colors representing letters). But the inherent ethos of the grammatical structures should be represented in a similar fashion visually. For instance, a period represents a hard stop, the completion of a thought. It is the end of the line, both grammatically and visually. A comma is more of a pause. So does that make it a space in the line? A curve? A change in weight?

[Read more]

After a meeting with Dana and a little down time to reconsider what direction this going in, I’m shifting the focus of my project. I’m flushing some ideas out here, but anything worth keeping will be moved to the Narratopolis project page.

I’ve spent most of today working on the base code to import and analyze the text for my project, and it’s promising to be a little more complicated than I’d hoped. I was thinking that I could just output a number of variables today and get to work on assigning them to various aspects of the visualization tomorrow, but given the relational complexity of the data, it looks like I’ll need to do a little more brainstorming about the cross-sections of info I want. For instance, I can get the number of words in the entire piece, the number of words in each paragraph, or each sentence, but storing the dat for all three of those is fairly confusing. If I could write to a database, it might be a little easier, but I’m not learning that by Wednesday.

On a broader scope, I’ve been considering the implications of combining the two projects — ICM and Metaforms — into a single idea with slightly divergent outputs. ICM is fairly straightforward, although may include a slightly more complex visualization at the end of the day. Metaforms, however, is looking like it will require a more developed rational for the project. The loose idea at this point is to take the narrative visualization that I’m working on for ICM and translate that into a physical form through the 3D Rapid Prototyping Service. But the more I think about it, the more I think that moving what I’m imagining as a narrative visualization into a physical form may be damn near impossible. I should be able to expand the idea to the z-axis with OpenGL or some other 3D processing library, but the form may be so complex, so intertwined and labyrinthine, that the RPS might not be able to make it. Not to mention that I’ll have to translate it into Maya, which I’m not even sure is possible. (Although this looks promising.) 

[Read more]