concrete poetry style shaped text

[MxAmperzand]

I haven't done anything like this before, but this seems pretty fun so here goes:

Idea: Making a book where the pages are concrete-poetry-style shaped text. I care less about coherency and more about making something that looks cool visually, though I am going to try to give it a storyline.

Intended process:

• Use concrete noun as topic for a single page (or possibly two page spread)
• Find/make silhouette image of that noun
• Pixelate the image into a grid for monospace characters
• Pull/generate lines of words with the correct character length to fit that grid

To be determined:

• Framing story/reason all the words are in shapes, and relatedly what shapes they are current possibilities:
  • video game journal that gives lore on all the items being picked up by the character, by possibly either a helpful tutorial character, the player character narrating their finds, or
  • horror journal entries trying to tell the character not to pick up these cursed objects, possibly because they trap you inside
  • mystery describing a bunch of items having gone missing/being stolen (in which case the text would be in the negative space around the empty object)
• How I'll shape the images (Do I allow single character pixels? Do I try to find a perfectly square monospace font or allow for tall pixels? All caps? Do I draw the pixel art or generate from images or find some online?)
• Whether to use jQuery/javascript or python (I'm more recently familiar with jQuery and will probably use it if I decide to pull lines from somewhere via APIs, but python's probably better for generatively pixelating images or outputting in PDF)
• What format to output the book as

[nmifsud]

I like all your framing ideas. The mystery one seems particularly promising. It could be cool to describe items in an oblique way, so that the shape of the text is necessary to comprehend the meaning (rather than being merely decorative). One approach might be to seed generated text with Wikipedia articles that are somehow stripped of explicit references to their subject.

In case it helps, I've made a simple thing related to your intended process.

Anyway, wherever you head with this, I look forward to seeing the end result.

[MxAmperzand]

Ooh, I do like the idea of not mentioning the actual object in it, that may have to be what I go with. Maybe a detective taking witness statements from distraught victims. I might have fun doing some bit of random generation of witness descriptions too. So far I think I'm going to make this in colab since I wanted to try it out, and I'm liking https://strlen.com/square/ as a perfectly square monospace font to use. I'll probably pull the images from 54px icons or something.

New problems to figure out:

• What pdf library to use
• how to pull correct length phrases from a text