preserving digital culture: tools & strategies for building web archives : tools and strategies...

30
Preserving Digital Culture: Tools & Strategies for Building Web Archives : Tools and Strategies for Building Web Archives Internet Librarian 2009 Tracy Seneca – California Digital Library

Upload: eric-evans

Post on 31-Dec-2015

216 views

Category:

Documents


1 download

TRANSCRIPT

Preserving Digital Culture: Tools & Strategies for Building Web

Archives : Tools and Strategies for Building Web Archives

Internet Librarian 2009Tracy Seneca – California Digital Library

Session Outline• Web content – why we’re building archives• Web crawling - under the hood• Tools available• Web Archiving Service Demo• Solutions for tough times - collaboration

Quick Tour

Web Content – Why Build Archives

Subject Area Author Sample DatesSample

Size Half-Life

Information Science Goh & Ng 1997-2003 2,516 5 years

Computer Science Spinellis 1995-1999 4,375 4 years

Law Rumsey 1997-2001 3,406 4 years

Medicine Veronin 1998-1999 184 3 years

• 2003: International Internet Preservation Consortium

• 2005: NDIIPP funds Web-at-Risk, Web Archives Workbench

Threats to Web Content

• Delivery model (many people access one copy)

• Site redesigns• Normal maintenance • Political change

– change of administration– policy changes

• Format

Researcher’s Perspective

• Study the topic / event• Study site change or web-based communication• Create stable citations for publications• Locate archived documents via catalog• Treat archive as a data set

What Makes an Archive

• Collection development – site selection

• Capture – harvesting content

• Curation – description and QA

• Publication – end user access

Archive Types

• Topical• Event• Domain• Document• Personal

Under the Hood

• Heritrix crawler• NutchWAX indexer• Open Source Wayback viewer

Open source tools from the Internet Archive

The Crawler1. Where do I start?2. Can I find that URL?3. Is there a robots.txt?4. What do I need to render that page? (CSS, graphics)5. What links can I find?6. Do those links fit the rules I was given?7. Do I have a flash / PDF / javascript file?8. Does that file have any links?9. For every link that fits the rules, start over!10. Keep going until I can’t find any more links or I hit my

time limit.

What the Crawler Spits Out

• ARC / WARC files– All of the content lumped together in large files– Keeps the archive simple and manageable – Need special tools to search and display

• NutchWAX• Open Source Wayback

• Massive amounts of content!

Why Should I Care?

• When you navigate a web archive, you’re interacting with a very different file structure

• These tools are constantly improving– Crawler gets better at capturing– Indexer gets better at ranking & scaling

• The Web is constantly changing– New technologies, new obstacles

Tools Available: Considerations

• Hosted vs. local

• Cost

• Public access

• Discovery / search options

• Capture configuration

• QA / Analysis Tools

• Metadata options

• Training & Support

• Ease of use

• Limits to:– Users– Archives– Sites– Storage

• Data Transfer

• Data Configuration

• Collaboration

• Rights management

Tools Available

• Hosted– Archive-It– Web Archiving Service– OCLC Web Harvester /

CONTENTdm

– Hanzo Web

• Local Installation– Web Curator Tool– CONTENTdm– NetArchive Suite

Archive-It• Hosted by the Internet Archive

• User-friendly interface, documentation, training

• Capture target = entire collection

• Public access automatic

• Dublin core metadata at seed level

• Limits = storage, # collections, # seeds

• Search full text, not metadata

• Highlight: “Scope It”

http://webarchive.jira.com/wiki/display/ARIH/Welcome

Web Curator Tool• Developed by National Library of New Zealand with input from the British Library

and other IIPC members

• User-friendly interface, strong user documentation for both technical staff and curators

• Rights management module

• Basic capture settings offered with access to all settings if needed

• Assumes a strong division of labor / specific order of events

• Capture target is flexible (sites or groups of sites)

• Dublin Core metadata

• Highlight: “Prune” tool

http://webcurator.sourceforge.net/

Web Archiving Service• Hosted by the California Digital Library

• User-friendly interface, documentation, training

• Capture target = site (flexible capture settings)

• Public access (optional)

• Some rights management features

• Limits = storage

• Search full text, not metadata

• Highlight: “show me all the new PDF files”

http://was.cdlib.org

• Web-based demos• User guides

Web Harvester / CONTENTdm

• Harvester hosted by OCLC• Access either hosted or local• Flexible metadata• Search metadata, not full text (except PDF)• Same public access interface as CONTENTdm

NetArchive Suite

• In use at Danish Royal Library 2004• OS release 2007• Tools developed for large scale and comprehensive

domain capture• High degree of control over crawlers• High degree of in-house expertise required• Documentation targets technical staff, not curators

• Highlight: QA tool that lets you click to grab missing images, files

Why have curatorial tools?

Web Archiving Service Demo

Rights Issues: Section 108 Study Group

• No advance permission needed to capture freely available web content

• “Freely available” = no login / fee

• Content owners can prevent capture via robots.txt and may request take down– Except government agencies

• Embargo period observed before archives are published

Large Scale Collaboration

• International Internet Preservation Consortium– Improving capture & display tools– Beginning registry of archives

• APIs to allow searches against different archives, no matter which archiving tool was used

End-of-Term Harvest

• Library of Congress, Internet Archive, California Digital Library, University of North Texas, GPO

• Nomination tool for managing 3000+ URLs for government agency sites

• Captures run at 4 institutions• Content replicated by partner institutions• Public access via Internet Archive

State of California Government Web Archive

Collaboration between

• State agencies/site owners and libraries• Across libraries• Librarians and faculty• Individual researchers

Questions?

[email protected]

• http://was.cdlib.org