ECDL 2006 - Session 10 - Next Generation Million Book Digital Libraries

Tuesday September 19, 2006
17:30 "Beyond Digital Incunabula: Modeling the Next Generation of Digital Libraries". Gregory Crane

This was an interesting presentation about the many ways in which we could mobilize digitized books - analyze and link the full-text in many ways.  A digital book should become much more than just a static PDF online - it should participate actively in a network of information.

http://www.perseus.tufts.edu/

http://dlib.anu.edu.au/dlib/march06/crane/03crane.html

http://ase.tufts.edu/faculty-guide/faculty.asp?id=gcrane

Separation of Content and Presentation

* extract chunks via XML

Recombinant Data

* Disassemble documents into pieces
* Recombine them on the fly

Dynamic Data

Books Talking to Each Other

Hybrid entty

Human/Machine/Services

Automatic Processes

e.g. Named Entity Analysis - figure out context of references to "Washington"

Lexical Analysis

* doing analysis of mapping between language and its translations

New User Interactions

* Readers talk, books listen

* Personalization

Million Book Libraries

* Google Books
* Open Content Alliance
* i2010 - in planning

Compared to curated

* 10 times bigger
* 10 times more noise
* etc.

Technologies and Domains

* Three core technologies
  - page image to text
  - text to data
  - one language to another

Million Books Workshop (to be announced)

Boston USA
May 22-24, 2007