It turns out that Amazon already does some (very basic) analysis of crowdsourced interpretation. They publish several lists of most popular highlights from Kindle readers.
Apparently this group of readers really thinks the first sentence of Pride and Prejudice is significant for one reason or another. But far more find something highlight-worthy about this line from Catching Fire (The Second Book of The Hunger Games):
Because sometimes things happen to people and they’re not equipped to deal with them.</blockquote>
This is a bit like Prism with tens of thousands if not millions of users (the quote above was highlighted by 17784 users at this time of this post) and no constraints on the meaning of highlights—no “categories.” The parallel was noted at least in passing by last year’s team. Kindle has come up a number of times with the current team so far, and most recently in Katina’s post this week about the social aspects of reading. Perhaps it’s time to revisit the comparison?
It would be interesting to try out some of the Prism visualization ideas on such a large pool of annotations. We could also experiment with computational linguistics techniques to make sense out of free text comments attached to highlights by users (it’s been proposed that Prism could use free text input too instead of fixed categories). But from what I can tell there’s no API that would allow us to work with the data. The closest I’ve found is a tool developed to scrape a single user’s highlights from the web. But one user wouldn’t do us much good. Access to the entire pool of public highlights is what would make this really interesting.
Another way we could think about overlap with Prism is in the reader software itself. How would this project change if we considered developing Prism as a plugin for existing software already in use by futuristic readers such as ourselves? Is there functionality in these apps/platforms that could benefit Prism?
Adobe Reader, for instance, has a wider palette of annotation tools (not just highlighting but free text comments, ovals, translucent boxes, arrows, etc.) that could lead to different user experiences and also different visualizations. Proposal: next week let’s try some annotation games with the Praxis team, but using PDFs instead of transparencies. E-mail them to me and I can separate the highlights and overlay them in Illustrator in a few minutes… An automated tool to merge highlights on PDFs would (I think) be very doable, and could even incorporate some Prismy visualizations.