Primo at the University of Amsterdam - Technology vs. Real Life
1. University library
Primo at the University of Amsterdam
Technology vs Real Life
Lukas Koster - Library Systems Coordinator - Library of the University of Amsterdam
@lukask – l.koster@uva.nl
EMTACL 2012, Trondheim, October 1-3, 2012
2. University library
Agenda
1-Discovery tools
1a-Technology
3-Indexing
4-User interface
2-Content
http://lib.uva.nl
Primo at the University of Amsterdam -
EMTACL12
3. University library
Technology vs Real Life
http://www.flickr.com/photos/brewbooks/3318600273
Primo at the University of Amsterdam -
EMTACL12
http://www.flickr.com/photos/35669523@N04/3310187686
4. University library
Discovery tools
A one-stop single-search-box web based solution for searching, browsing, discovery and
delivery of print and digital publications and objects available from library
collections, institutional resources and academic publishers, using a unified index
Content
• Print publications User interface
• Digital publications • One stop
• Objects • Single search box
Indexing
• Library collections • Web based
• Unified index
• Institutional resources • Searching
• Academic publishers • Browsing
• Discovery
• Delivery
Primo at the University of Amsterdam -
EMTACL12
5. University library
Discovery tools
What I would like:
Gateways to all information a library has access to
Primo at the University of Amsterdam -
EMTACL12
6. University library
Discovery tools: the environment
Academic libraries
Teaching
Research
Access
Subscriptions
Freely available
Information
Traditional publications
All other types
Primo at the University of Amsterdam -
EMTACL12
http://commons.wikimedia.org/wiki/File:Graduation_hat.svg
8. University library
http://www.flickr.com/photos/manchesterlibrary/2034771121
Technology
Harvesting & Indexing
http://www.flickr.com/photos/quiltsalad/5991773081/ Primo at the University of Amsterdam -
EMTACL12
9. University library
Technology
Discovery
frontend
User interface
Central
Metadata
Index
Indexing
Harvesting
ejournalejournal ILS
Image
ejournal database
UI
External External Local
External
database database Repository database
database UI
UI Primo at the University of Amsterdam -
Content UI EMTACL12
10. University library
Technology
Discovery
frontend
User interface
Central
Metadata
Index
Harvested and Indexed Content
Primo at the University of Amsterdam -
EMTACL12
11. University library
Content
Primo at the University of Amsterdam -
http://www.flickr.com/photos/mollyblock/7941237158 EMTACL12
12. University library
Content
Theoretically (technically)
we can harvest everything
we have access to
Primo at the University of Amsterdam -
EMTACL12
13. University library
Content
Discovery
frontend
User interface
Central
Metadata
Index
Indexing
Harvesting
ejournalejournal ILS
Image
ejournal database
UI
External External Local
External
database database Repository database
database UI
UI Primo at the University of Amsterdam -
Content UI EMTACL12
14. University library
Content
Discovery
frontend
User interface
Shared
Metadata Local
Index Metadata
Index
Indexing
Harvesting
ejournalejournal ILS
Image
ejournal External External Local
database
External
database database Repository database
database
Primo at the University of Amsterdam -
Content EMTACL12
15. University library
Content
Discovery
frontend
User interface
Shared
Metadata Local
Index Metadata
Index
Harvested and Indexed Content
Primo at the University of Amsterdam -
EMTACL12
16. University library
Content
Local
Shared Local
Metadata Metadata
Index Index
Primo at the University of Amsterdam -
EMTACL12
17. University library
Content
Local
Shared Local
Metadata Metadata
Index Index
ejournalejournal ILS
ejournal External External Local
External Repositories SFX
database database
database
Primo at the University of Amsterdam -
EMTACL12
18. University library
Content
Local
Shared Local
Metadata Metadata
Index Index
ejournalejournal ILS
Student Theses
ejournal External External Local
External Repositories SFX
database database
database
Content
provider Content
provider Content
provider
System
vendor
Primo at the University of Amsterdam -
EMTACL12
20. University library
Content
Content types
Mostly: Traditional publications
Books
Articles
Also other types
Datasets
Maps
etc.
Primo at the University of Amsterdam -
EMTACL12
21. University library
Content
http://www.flickr.com/photos/adactio/2144119569
In reality
we can’t harvest everything
we have access to
Primo at the University of Amsterdam -
EMTACL12
22. University library
Indexing
Primo at the University of Amsterdam -
EMTACL12
23. University library
Indexing
Theoretically (technically)
we can index everything
unambiguously
Primo at the University of Amsterdam -
EMTACL12
24. University library
Indexing
Local
Shared Local
Metadata Metadata
Index Index
Two separate indexes
Primo at the University of Amsterdam -
EMTACL12
25. University library
PNX
Indexing <search>
<author>
<title>
</search>
<display>
<author>
<title>
</display>
Or similar,
<facets>
<author>
etc.
Normalising
Harvesting <type>
<date>
Source </facets>
Data records PNX <links>
source PNX
PNX </links>
PNX <delivery>
</delivery>
Primo at the University of Amsterdam -
EMTACL12
26. University library
Indexing
Local
Shared Local
Metadata Metadata
Index Index
No deduplication across indexes
No FRBRisation across indexes
Primo at the University of Amsterdam -
EMTACL12
27. University library
Indexing
Local
Shared Local
Metadata Metadata
Index Index
Consolidate both indexes
Adapt local indexing to shared indexing
Primo at the University of Amsterdam -
EMTACL12
28. University library
Indexing
<creatorcontrib>
Author names Beckett, Samuel
</creatorcontrib>
<creatorcontrib>
Samuel Beckett 1906-1989
</creatorcontrib>
<creatorcontrib>
Beckett, S
</creatorcontrib>
Works reasonably well <creatorcontrib>
Samuel Beckett
Multiple search variants </creatorcontrib>
But: strings, no unique identifiers
Primo at the University of Amsterdam -
EMTACL12
29. University library
Indexing
Topics
Again: strings, no unique identifiers
Subjects/topics/keywords etc. are
taken from each datasource ‘as is’
Primo at the University of Amsterdam -
EMTACL12
30. University library
Indexing
Resource types
Match Resource Types codes across indexes
One Resource Type per record
Primo at the University of Amsterdam -
EMTACL12
31. University library
Indexing
Resource types
Interesting example from institutional repository
MODS/DIDL
<typeOfResource>text</typeOfResource>
<genre>info:eu-repo/semantics/doctoralThesis</genre>
Primo at the University of Amsterdam -
EMTACL12
32. University library
Aside: Primo “hackable”
Ex Libris Open APIs
customisable, plugins, addons
Linked Open Data Special Interest Working Group
http://igelu.org/special-interests/lod
Primo at the University of Amsterdam -
EMTACL12
33. http://www.flickr.com/photos/maveric2003/3822708724/
University library
Indexing
In reality
we can’t index everything
unambiguously
Primo at the University of Amsterdam -
EMTACL12
http://www.flickr.com/photos/profzucker/3754015526/
34. University library
User interface
Primo at the University of Amsterdam -
EMTACL12
http://www.flickr.com/photos/mafleen/125422650
35. University library
User interface
Theoretically (technically)
we can find all we need
with one search
Primo at the University of Amsterdam -
EMTACL12
37. University library
User interface
Setting context
Before Search After
Advanced search, etc. Refine results/facets
Subject Subject
Discipline Discipline
Scope Source
Type Type
Date Date
… …
Primo at the University of Amsterdam -
EMTACL12
38. University library
User interface
Broad Refine
Facets
http://www.flickr.com/photos/eirasi/2084477067/
Primo at the University of Amsterdam -
EMTACL12
40. University library
User interface
Requires uniform classification
by subject of each item
In Local and Shared index
Discipline At the moment only available
for relevance ranking in
On Journal level ScholarRank
Primo at the University of Amsterdam -
EMTACL12
41. University library
User interface
No discovery desired!
Search on Title , Title + Author
Known item
Primo at the University of Amsterdam -
EMTACL12
42. University library
User interface
Different audiences, context
Depending on
context, different search
interfaces may be
appropriate
http://www.flickr.com/photos/rrrrred/3923807023
Primo at the University of Amsterdam -
EMTACL12
44. University library
Technology vs Real Life
We can’t harvest everything
We can’t index unambiguously
We can’t find all we need with one search
YET!
Primo at the University of Amsterdam -
EMTACL12
45. University library
YET!
http://www.flickr.com/photos/katerha/7071545621
NEXT?
Primo at the University of Amsterdam -
EMTACL12
Notas del editor
Marketing speak
Not focusing on discovery
Not focusing on discovery
Fast, uniform
Content is harvested and indexed in a consistent uniform way
Content is harvested and indexed in a consistent uniform way