Idea: Describe all the things
(This is an idea for a future readmspa.org project.)
The idea is to generate textual descriptions of every image and Flash.
(I currently have transcripts of all the words from images and Flashes, but apart from in interactive Flashes, there are no descriptions of anything else in them. The game transcripts are done as walkthroughs. The instructional parts don’t count towards the word counts though.)
I have three main reasons to do this.
- Contemporary Android and iOS mobile devices don’t play Flash. (Advanced users can unlock and/or install special software, but regular users won’t.) These days, there are more of these devices than there are PCs! Having text that describes each Flash preserves the Homestuck story by making it more “readable” on those devices.
- Descriptions for all images as well would further mean that blind readers could enjoy Homestuck more.
- Even though it is comparing apples to oranges, I want to measure the narrative weight of the non-textual parts of Homestuck, so as to compare its length more fairly with other stories. The original idea is due to notEgbert. He came up with a crude but useable formula in his original Measuring Homestuck report, and I’ve rolled it into my rolling statistics. But the only true measure of how many words a picture (or animation) is worth is to actually describe it.
Cost:
Huge! There’s no real way to do it without involving lots of…
Other people:
This would have to be a crowd-sourced effort. Some consensus on style and appropriate level of detail would be desirable, and I’d be happy to provide tools and organisational infrastructure, but probably being open to anything is the best approach to build critical mass.
I did a tiny experiment on this once before. It worked OK so far as it went.
To just achieve the a better measurement of narrative weight, I note it isn’t necessary to describe everything - just to describe enough samples of varying complexity to get an idea of typical weight, then weight all the other things by classifying them relative to the samples.
Callbacks:
Descriptions of images and Flashes would mean you could find them using a proper search.
