|::::::::::::::::||NYU | Tisch School | ITP|
Final Project Proposal
You could tell a lot about people's personal bookshelves by the king of books they are reading, where the book is (specifically the bottom shelf or on their bedside table), how they organize their shelf (prioritizing more frequently or less frequently read books).
Two personal "book recommendation" examples come to my mind:
1) When I used to date my partners, I would always see what books we had in common to see if we were close in personalities.
2) In March 2006, I would notice many people (mostly strangers) with the Malcolm Gladwell book Blink. My first observation of someone reading this book was a woman at Starbucks. She had a pile of papers from what I assumed a class she taught. The first paper in her pile had a name and "Psych" label, which also told me that her expertise was in Psychology. I then saw a man reading this book on the subway. My curiosity led me to the public library. When I did a search for this book, there were 252 people waiting to read one of two copies in circulation. So I finally purchased this book online.
We would like to reproduce this experience in our networked objects project. We examined two spaces, the physical and the digital. The physical spaces include public libraries, academic libraries and book stores. These spaces are good at locating a book physically, but do not recommend books. Some have complicated interfaces, such as call numbers which rarely relate to the book. The digital spaces we observed are web sites like Amazon, Douban, Delicious Monster, Bill Monk (just tracks, doesn't recommend) and the Library Thing. Delicious Monster was the closest site with some physical component, but all these sites lack an accurate recommendation system. Amazon is good at recommending books to people who have yet to purchase the book, and Douban is good at recommending books to people who have the same books. None of these digital spaces provide the physical content of the book. We propose that our smart shelf will bridge the physical and digital spaces.
Three User Scenarios
I. Personal Use - Smart Shelf on a personal level will allow the user to create several reading circles, and these connections will be denser.
II. Public Use
III. Public Use - In a Community Setting, Library in Japanese Room
Personal Use. Each book that a user wants to share with the circle will have one unigue RFID tag. That information will be stored on our database along with the ISBN number of the book. That ISBN number is then matched with other copies of the exact same book. Every incidence the book is taken off the shelf, it will ping our database, and then depending on whether members of the circle want to participate, it will log that information in a folder in their account, and if they wish to directly contact this person to start dialogue, they may.
Public Use. Currently, random RFID tags are used in book stores for security purposes. The books have tags, which are on, until the customer purchases the books. The book is then scanned to turn off the tag, so the alarm isn't set off when the customer leaves the store. RFID tags can be entered in our database to count every incidence a book is taken off the shelf. Depending on whether the shelves have LED lights or screens, the display will notify that the customer the most "popular book." This information can be sorted. For instance, Tim O'Reilly's book on the subject of AJAX, may come up to the top of the list for web development tool books. This information may be juxtaposed with other book recommendations scraped from the Amazon site (e.g. data visualization of book geneology for Malcolm Gladwell's Tipping Point via back and front citations). Citations also connote a similar type of networked circle, the author's circle.
Community Use. Every book is tagged with RFID, which is also linked to our database. The user must also have a library card, which is assigned to an individual. This book's history is tracked, which may include the user's personal notes. The display will show the history of users for that one particular book. This way, the current user could determine if the book at hand will suit the user's need.