This session was, without a doubt, my
favorite session of the day, and may wind up being my favorite
session of the conference, if for no other reason then the incredibly
impressive technology that was described.
The session began unassumingly enough
with Chad Mairn, CTO of Novare Library Services and librarian for
West St. Petersburg Community Library providing a brief overview of
how libraries are in the business of helping people discover new
books. He pointed how West St. Petersburg has a poster with some new
books and a QR code that directs a user to the complete list. He
then talked about good sources for DRM-free ebooks like
drmfree.calibre-ebook.com, Project Gutenberg, OpenLibrary.org, and
Overdrive (which has both DRMed and DRM free books). He mentioned
an interesting idea about putting up a sign in BestBuy to direct
people who had just purchased ereaders to the library for books and
offering free workshops for the public to help people with their
devices. He also talked about how libraries are in a position to
work with authors who want to self-publish their materials.
Then Chad discussed some of the book
discovery services out there like Library Thing, Good Reads, BookLamp
(which I wasn't familiar with), and Small Demons (which I also was
not familiar with) and capped that off with the genres page at
WorldCat (www.worldcat.org/genres/)
which was interesting.
After covering this in probably about
10 minutes, Chad invited Aaron Stanton from BookLamp to come up and
talk about his website, which is about when the session became really
interesting.
Aaron mentioned that most of the
services that Chad had just mentioned are social book discovery
tools. They have limitations because certain books are more widely
read and more favored, so they float to the top of the services.
This isn't necessarily because they are better, but just because they
are more popular. There are many books below that may be of great
interest to readers but are just not noticeable because they haven't
had enough readers and they are kind of doomed to that state in a
social discovery service.
This is where BookLamp differs from
those.
Rather than working on a social
recommendation service, Aaron (who from his presentation is clearly
an avid reader) decided to develop a way of analyzing the content of
a book with computer algorithms. Using some apparently incredibly
sophisticated code, Aaron and his team have analyzed the content of
books and can identify features that one would think would be nearly
impossible for a computer to identify such as pacing, writing style,
action, characters, and themes. This information can then be put
into a two or three dimensional graph to give the book any number of
unique, identifying fingerprints. This whole initiative is called
the Book Genome Project, and of the 90 some thousand books the Book
Genome Project has processed, 20 thousand are on BookLamp, with more
to come.
It sounds like a really interesting
project and something that should be a useful readers advisory tool
and book finding aid well into the future.
No comments:
Post a Comment