Digital Libraries: Content, Discovery, Interoperability, and User Experience, Study Guides, Projects, Research of Computer Science

Download Digital Libraries: Content, Discovery, Interoperability, and User Experience and more Study Guides, Projects, Research Computer Science in PDF only on Docsity!

Digital libraries

Brian F. Cooper

October 28, 2003

Overview

„ What is a digital library?

„ Content

„ Information discovery

„ Interoperability

„ User interaction

„ Security

„ Preservation

What is a digital library?

„ What is a library?

„ What is a digital library?

Digital libraries

„ National Digital Library (Library of Congress)

„ Digital Library of Georgia

„ Alexandria Digital Library

„ Perseus Digital Library

„ Internet Archive

„ ACM Digital Library

„ Georgia Tech Library Digital Initiative

„ Citeseer

„ Google

Content

„ How much content is there?

† Audio collection of the works of Beethoven – 20 GB † Library floor of books on shelves – 50 GB † Academic research library – 2 TB † Printed collection of the Library of Congress – 10 TB † 3 years of EOS data – 1 petabyte † All printed material – 200 petabytes † All words ever spoken by human beings – 5 exabytes

Source: http://www.jamesshuggins.com/h/tek1/how_big.htm

Content

„ Where does the content come from

† Digitized

† Born digital

Content

„ Types of content

† Documents

† Images

† Audio/video

† Geo-spatial

† Searchable databases

Geospatial content

„ Alexandria Digital Library Project

† Commercial example: Mapquest

„ Georeferenced images

† Maps † Satellite images

„ Gazetteer

† Find geographic entities based on name † Like the index in an atlas

Source: http://www.alexandria.ucsb.edu

Use the librarian

„ Idea: things are organized well

„ But…

† We make data faster than librarians can organize it † (Is this a good thing?)

Use the “card catalog”

„ Search metadata

Author: William Shakespeare Title: Hamlet Year: 1601 Subject: Depressed Danish princes

„ Search content

“question outrageous fortune”

To be, or not to be: that is the question: Whether 'tis nobler in the mind to suffer The slings and arrows of outrageous fortune, Or to take arms against a sea of troubles, And by opposing end them? To die: to sleep; No more;

Metadata

„ Issues

† Where does it come from?

† What does it mean? „ E.g. “author” versus “creator” versus “artist”

† How do you manage it?

† What happens over time?

Metadata

„ Example: Dublin core

„ Core elements

† Title, Creator, Subject, Description, Publisher, Contributor, Date, Type, Format, Identifier, Source, Language, Relation, Coverage, Rights

„ Other elements

† Audience, abstract, replaces … † 33 elements † Can be extended

Source: http://dublincore.org/documents/2003/03/04/dcmi-terms/

Google

„ Basic idea: PageRank

† “A page is important if important pages point to it”

Stanford University

Bob’s Homepage

The ugly math

„ The PageRank equation

PR(A) = (1-d) + d (PR(T1)/C(T1) + ... +PR(Tn)/C(Tn))

† PR(A) – the PageRank of a page † PR(Ti) – the PageRank of a page pointing to A † C(Ti) – the number of outgoing links of Ti † d – Dampening factor so the computation converges

„ The theoretical justification † The “random surfer model”

Source: S. Brin and L. Page, "The Anatomy of a Large-Scale Hypertextual Web Search Engine," Proc. 1998 WWW Conf., 1998

Personalized search

„ Find information that is relevant to you

† Classic problem: “apple” versus “Apple” † Idea: pages you like are “important” † Calculate page rank for other important pages

„ Challenges

† This is expensive! † Calculate PageRank over 3 billion pages for each user

The ugly truth

„ Google can be “gamed”

† Create lots of links to your page † Weblogs

„ PageRank little used anymore

† Anchor text, word similarity, other secret things

„ Google is “broken”

Interoperability issues

„ Legacy systems

„ Data schema

„ Data encoding

„ Data semantics „ Query language

„ Result ranking

„ Limited interfaces

„ Current protocols solve only part of the problem † TCP, HTTP, SOAP, etc.

DL-specific protocols

„ Z39.

† Search and retrieve “card catalog” information

„ SDLIP/STARTS/SDARTS

† Limited state search, retrieval and metadata

„ OAI

† Protocols and toolkit for “Open Archives”

User interaction

„ How can a user get the information?

† The old days

Things have changed

Things have changed

Stanford PowerBrowser

„ Browsing on PDA’s

† Limited screen real-estate † Limited PDA processing † Limited bandwidth

„ Solution

† Innovative user interface † PDA/proxy interaction

Source: http://www-diglib.stanford.edu/~testbed/doc2/PowerBrowsing/desc.html

Browsing on mobile devices

„ Stanford PowerBrowser

Browsing on mobile devices

„ Stanford PowerBrowser

Browsing on mobile devices

„ Stanford PowerBrowser

Browsing image libraries

„ Stanford PhotoBrowser

† How to organize boxes of photos? † Digital photos have identifying information „ Date „ Maybe location

Source: Graham, Garcia-Molina, Paepcke and Winograd. "Time as Essence for Photo Browsing Through Personal Digital Libraries.” JCDL 2002

PhotoBrowser

PhotoBrowser