Notes
Slide Show
Outline
1
OAI User Services
  • Kat Hagedorn, UM
  • University of Michigan
  • 11/10/2005
2
Ready, get set,…
  • You have your metadata ready…
  • You have your tools for uploading…
  • You’re officially a data provider…
  • So, how does the metadata get used once it’s available through OAI?
3
OAIster
  • Contains “all” OAI records
    • collects only records that point to digital objects
    • but does harvest all data providers
  • http://www.oaister.org/
  • 549 institutions; almost 6 million records
  • 37% US, 63% international
  • 16% eprints, 11% DSpace, 4% ContentDM, 3% DigitalCommons
4
MODS Portal
  • DLF members only (development is part of DLF IMLS grant)
  • MODS metadata records only
  • http://www.hti.umich.edu/m/mods/
  • 4 institutions (LoC, Indiana, OCLC, Univ of Chicago)
  • Over 330K records (mostly LoC)
5
DLF Portal
  • Like the MODS portal, but not specific to MODS
  • http://www.hti.umich.edu/cgi/b/bib/bib-idx?c=imls;page=simple
  • Simple DC records
  • 43 institutions; over 880K records
6
Other harvesters/portals
  • Format-specific, ex. Sheet Music Portal
  • http://digital.library.ucla.edu/sheetmusic/
  • Country-specific, ex. Cyberthèses
  • http://cybertheses.francophonie.org/archives.php
  • Software-specific, ex. Eprints.org, PKP
  • http://www.eprints.org/software/archives/
  • http://pkp.sfu.ca/harvester/archives.php
7
UM system
  • First three portals all built at UM
  • Developed a system for
    • harvesting records (DC, now also MODS)
    • transforming/normalizing them
    • ingesting them into DLXS Bibliographic Class
    • search and display of records
  • Use DLXS: digital library creation software
  • Built our own harvester (in perl)
8
System design
9
Your data in our system
  • MODS Ô DLXS BibClass
  • before…
10
Your data in our system
  • MODS Ô DLXS BibClass
  • during, phase one…
11
Your data in our system
  • MODS Ô DLXS BibClass
  • during, phase two…
12
Your data in our system
  • MODS Ô DLXS BibClass
  • after…
13
Your data in our system
  • MODS Ô DLXS BibClass
  • display…
14
Evidence of use
  • Articles, both scholarly and otherwise
  • Users write about it on blogs
  • Data providers care enough to complain J
  • User stats
    • for OAIster, regularly in 18-19K+ hits/day range
    • hundreds of thousands of hits/day on Yahoo!
15
Pitfalls…
  • Complex data gets “squashed” into simpler, flatter bibliographic data format
  • Especially for MODS
  • Working on appropriate ingest into Bibliographic Class so complex MODS elements are better reflected
16
More pitfalls…
  • Don’t know who users are
  • Tested 3 years ago, for interface issues
  • Need to test functionality with current grant’s Scholar’s Panel, end-users, and…
  • Current and potential data providers, such as yourselves
17
Next steps
  • Lots of things planned for all portals…
  • MODS/MARC integration
  • Thumbnail grabber: include thumbnails in results, as in CIC Portal
  • Date normalization
  • Make metadata downloadable from portal, and not just as DC or native XML
  • Clustering for better search/browse
18
CIC Portal
  • Over to Sarah… (no Caesar jokes this time!)
19
Questions
  • Kat Hagedorn
  • khage@umich.edu
  • University of Michigan
    • Digital Library Production Service
  • www.oaister.org
  • www.dlxs.org
  • www.umdl.umich.edu