how the web changes your reading habits 06.24.2005, 2:17 PM
posted by ben vershbow
An article in yesterday's Christian Science Monitor looks at two research projects currently underway in Palo Alto, California - one at Xerox PARC, the other at Stanford. Both are building tools and devising methods to improve online reading, albeit by different approaches. The PARC project is developing ScentHighlights, an "enhanced skimming" function based on keywords and the associative processes of the human brain. On paper, we highlight important passages, or attach sticky notes, to make them more readily retrievable later on when we're re-reading, studying, or compiling notes. The PARC researchers are taking this a few steps further, exploiting the unique properties (and addressing the unique challenges) of the online reading environment. With ScentHighlights, the computer observes what the reader is highlighting and selects other passages that it thinks might be relevant or useful:
We perform the conceptual highlighting by computing what conceptual keywords are related to each other via word co-occurrence and spreading activation. Spreading activation is a cognitive model developed in psychology to simulate how memory chunks and conceptual items are retrieved in our brain.
While the PARC team is focused on deepening the often fractured experience of reading online, where the amount of text is overwhelming, the Stanford project is experimenting with a method for sustained reading in an environment that can barely handle text at all: the tiny screens of cell phones and mobile devices. Using a technique called RSVP (Rapid Serial Visual Presentation), BuddyBuzz flashes words on the screen one at a time. It takes some getting used to, but apparently, readers can absorb up to 1,000 words per minute. Speed is adjustable, and the program is already set to make the tiny, natural pauses that come at commas and periods. The initial release of BuddyBuzz will syndicate stories from Reuters, CNET and a handful of popular blogs.
filologanoga on June 25, 2005 5:57 AM:
The first project reminds me uncomfortably of certain software trying to guess what I really want to write, or do. With the result that the guessing function gets turned off. Hopefully, the ScientHighlights algorithms will be more linguistically intelligent.