You can get insight into loosely connected conversations about books at All Consuming.
Think about all the conversations taking place on the Web at this moment. Unlike spoken conversation, conversations on the Web have a physical presence, a shape that persists over time in the form of digital text, and a visible trail left behind by people participating in the conversation. Trying to figure out what these conversations are about and following them from beginning to end, however, is not as easy as it might seem.
Conversations rarely give indicators allowing you to easily tell what they are about. Creating a script to determine this would require that it be able to interpret the meanings of sentences. To complicate things further, conversations can expand quickly to fill the entire community's consciousness overnight, or may lie dormant for months or years before being picked up again. Tracking and grouping these various bits of conversations has become a great opportunity and challenge for tool developers. All Consuming (http://allconsuming.net) is one web site that uses Amazon's friendly Web Services to attempt to track a tiny fraction of these conversations: those about books.
Tracking and following conversations about books became possible because these conversations had a couple fortunate characteristics.
First of all, books have unique identifiers. They have ISBNs (International Standard Book Numbers) that are recognized by almost everyone, including retailers, libraries, and, most importantly, the people having the conversations. These unique identifiers make it easy to find out when different people are talking about the same book.
Second, Amazon's Web Services make it easy get detailed book information about any ISBN. Even though people might occasionally misspell an author's name, link to a different version of its title, or leave the title or author out altogether, you can still easily figure out what book they're talking about if they link to Amazon or another retailer that has the ISBN in the URL.
Combining these two advantages, it took me just a few days to put together a web site that grouped conversations about the same book together from all over the Web. All Consuming slices and dices these conversations in a couple different ways.
Every hour, All Consuming visits the recently updated weblogs and looks for links to Amazon, Barnes & Noble, Book Sense, and even to itself. For each link that it finds, it checks to see if it had found that book on that web site before, and if not it saves it as a newly mentioned book. The newly mentioned books from each hour are listed at the top of All Consuming's home page, as you see in Figure 3-11.
In addition to the standard hourly report, there is also an hourly list of the books mentioned the most this week, giving preference to the most recent. When half a dozen web sites mention the same book on the same day, it will appear near the top of the list so you can instantly see when a book is generating buzz in the weblog community. You can also see snippets of the conversation on All Consuming or go directly to the site to see the conversation in context.
In addition to showing the most talked about books on the Web at the moment, there is a way to find out who's been talking about any book in Amazon's catalog. This allows you to visit a grouped conversation about a book even if it hasn't become active enough to make the front page.
Yet another way to slice the data is to display all of the books mentioned by a particular weblog or person. Every weblog has its own page on All Consuming, with a list of books mentioned on the weblog or directly through All Consuming.
If you want, you can also create an account at All Consuming, add books to your Currently Reading list, and then add a line of code to your site [Hack #48].