4. Kahn-Wilensky Web
Handle
I search
a paper
about...
Repository
Service
Repository
Provider
5. Hypothesis 1: One Network & Common Schema
Hypothesis 2: Object Oriented Design
Hypothesis 3: The Resource is the Message
6. Hypothesis 1:
One Network & Common Schema
OpenAIRE
Student
Papers
Google
Europeana
CRIS
PhD
Inst.
Repository
Digitized Data
Material Sets
7. Hypothesis 2:
Object Oriented Design
•(Complex) Digital Objects
• Unique identifiers
• Key-Metadata + 1 or more data streams
• Metadata schemes invariant over repositories
• Specialized data types + service bindings on these types
• Fragmented input
• Incomplete
• Metadata, Files, Metadata + Files
• Many schemas
• Simple data types (String, Array, Map) many native functions on these
types
8. Researcher
Hypothesis 3: The Resource is the Message
DNS
Dr Müller
Repository Google
ISI
Researcher
PubMed Homepage
11. Copy & Paste
Functional Style of Programming
on native hashes and arrays
“It is better to have 100 functions operate on one data structure
than to have 10 functions operate on 10 data structures.”
Alan J. Perlis
15. Store
Schemaless storage
Elastic Solr
Search
JSON
store
Mongo DBI
DB
$store->add({})
$store->search()
title:“krank” and subject.local:“cycle”
title any “krank” and subject.local any “cycle”
16. Import
Import from many sources
Atom LDAP DBI
JSON
importer
MARC OAI SRU
$store->each({})
$store->first()
$store->rest()
$store->select({})
$store->any({})
$store->many({})
26. Project Plan
• Catmandu: Open Source Data Toolkit
• LibreCat: Example Programs:
• LibreCat-Search, LibreCat-Citation,
LibreCat-Grim, LibreCat-Archive ...
• Suite of repository add-ons:
• Project Database, Research Groups,
Authority Files
27. • Nicolas Steenlant - Ghent
• Nicolas Franck - Ghent
• Patrick Hochstenbach - Ghent
• Snorri Briem - Lund
• Dave Sherohman - Lund
• Jörgen Eriksson - Lund Thanks!
• Maria Hedberg - Lund
• Friedrich Summann - Bielefeld
• Najko Jahn - Bielefeld
• Vitali Peil - Bielefeld
• Petra Kohorst - Bielefeld
http://librecat.org