It’d be cool to have special support for stereoscopic pairs
For the uninitiated, a stereoscopic pair, often referred to and tagged as a stereogram, is a pair of images representing slightly offset views of the same scene. When one image is presented to the left eye and the other to the right eye, the brain fuses the two images into a single, cohesive 3D scene. Stereoscopic pairs can be presented in a variety of formats for various viewing methods (mainly freeviewing, viewing with those red and blue glasses, and viewing in VR). Any given format can easily be produced from the original images.
At the moment, the format of the pair (and naturally, who is capable of viewing it,) is decided by the uploader.
The linked repository includes my implementation using the HTML canvas. It is functional but has significant room for improvement (see the readme for details). The implementation loads the two images from the folder in the webpage. Naturally, this would be swapped out for the method Derpibooru currently uses to obtain the image.
Note that the anaglyph generator requires reading from a canvas that, unless this webpage is served from an actual HTML server, will have been loaded from the client. This is disallowed for security reasons. If you want to try out the anaglyph generator without starting up an actual HTML server, it is possible to disable the security feature on chrome by starting up the browser from the command line with the
--disable-web-security option. You should never browse the internet with this feature disabled.
I’m not clear on the details of implementing this on the actual website, but ideally it should be about as simple as adding some html and/or js and putting a button linking to this page above posts. Possibly only above posts tagged as stereograms. More in-depth integrations could be done, but that would be the easiest.
It would really improve their accessibility and hopefully encourage this medium, which I think is a really cool one.