« contagious media: symptom of what's to come? | Main | one tree in a forest of information »
amazon: inching toward semantic 
04.13.2005, 5:08 PM
posted by ben vershbow
Sometime in the last few days, Amazon.com unveiled three new features for its Inside the Book search: "books on related topics," a "100 most frequently used" concordance (above is the concordance for Orality and Literacy by Walter J. Ong), and "text stats." The stats are pretty funny - in addition to page, word and character count, they measure a book's "complexity" as well as its "readability" according to three established indexes, including the famous and amusingly named "Fog Index" (as though it rated the density of mental fog between a reader and a book). It also includes so-called "fun stats" like words per dollar and words per ounce.
Some of these features seem a little trivial, but there's no denying that Amazon is moving surely and steadily toward a comprehensive semantic browsing system (other recent innovations are Statistically Improbable Phrases (SIPS) and Citations). Though still crude compared to what it might eventually become, you can begin to glimpse the pleasures and uses it will afford. Amazon can never replace the social and tactile pleasures of browsing a physical bookstore, but it's doing a good job at making the virtual bookstore a more exciting place.
Posted by ben vershbow on April 13, 2005 05:08 PM
tags: Libraries, Search and the Web
