smashingboxes / fraqture Goto Github PK

We would like to make two improvements to the tweetreader image:

• Instead of fading to the SB logo, it should draw in the photo attached to the text.
• It runs a bit long. Exit criteria should be added.

Not needed for NOEW (or perhaps at all)

The pixelate image starts too granular; it isn't until several rounds in (over a minute) that people understand what it is doing.

Right now, if a drawing writes outside of the 18 rows and 30 columns, it is possible to crash the application. Some bounding would be great.

We would like to make two minor changes:

• Use colors from the image when filling in
• Make the size of the squares being swapped roughly equal to the LED squares.

Three minor improvements for shifting grid:

• Remove pink outlines
• Slow it down
• Add a delay to the end

Right now, every sketch that uses an Arduino will attempt to connect to the Arduino during its setup phase. When we begin cycling these drawings, the second drawing will crash because the port will already be open. I'd like to make a serial interface that:

• Asks the user what port they would like to connect to upon opening the program. You can use list-ports from our serial library to enumerate their options.
• Opens the serial once during global set up, then puts the serial resource in the options hash being passed in to the sketches.
• Update all sketches that use the serial port to use this new interface.

The goal is to integrate a live camera feed into the Storefront Installation.

Features:

• includes a countdown
• captures a live image
• runs the image through one of the glitch algorithms
• transitions between live camera & seeded content

• When typing the message, use the bottom LEDs as a keyboard.
• When scrambling the message, scramble the LEDs as well.

A simple test suite could run each of the drawing commands (eg lein run drag) and kill it after 30 seconds or so, just to make sure each one runs without error. It could also look for stack-traces.

The tweetreader function needs an exit condition. I'd like for it to be 1-2 minutes after the textify animation starts, but am open to suggestions.

Also look into if we need to delete them because of file size.

Some glitches work better on logos, and others on rasters. If we split the folders in two, we can replace stream/get-image with stream/get-logo and stream/get-raster.

We currently have a mechanism for grabbing images for display, lets also have a way to grab text. Ideally we want to have some weighting system too to show more recent images.

Rework the drag drawing so that:

• Not all columns trickle at the same time.
• Columns trickle in from the top using the LEDs

We want to provide an animation when taking a picture, something along the lines of "3.. 2.. 1.. (flash all LEDs).

As a jumping off point for adding live content, we can first find some way to utilize our current infrastructure by having a module in Clojure that can take a picture and place it in the images folder or have Clojure call out into a script in some other language that does so.

Change the cycle drawing so that during the day it uses live camera feed + seeds content in the main glitch cycle, but at night it uses more subtle generative art.

Our readme has a couple of broken images and could be more descriptive and helpful.