The Digital Repository Infrastructure Vision for European Research

1

• The Digital Repository Infrastructure Vision for
  European Research
• in a repository ecology
• or
• Why DRIVER is not a pond!

2
Disclaimer Constructive Criticism

• Introductory Note
• it is very common in conceptual discussions
  that theories are presented together with their
  critics
• this contribution is meant as such an academic
  crtitique that shall help
• to further mature the ecology approach
• to enrich the discussion
• to point out strengths and problems
• to offer alternative approaches

3
Authors

• Muriel Foulonneau
• CNRS, Lyon-FR
• Paolo Manghi
• CNR-ISTI, PISA-IT
• Natalia Manola
• University of Athens, GR
• Wolfram Horstmann
• State and University Library Göttingen

PAPER
PRESENTATION
4
The setting

• The ecology approach suggests
• the inclusion of non-technical aspects in the
  frameworks (or models) for repository systems for
  the sake of easier comprehension
• the analogy of ecology as an explanatory
  framework for repository and service interactions

5
Terms

• Mapping of concepts in the repository ecology
  (Robertson et al.) to DRIVER

6
Scale

• Organism human
• Researcher, Developer, Administrator
• Population group of humans
• Institutional employees, conference visitors
• Community interacting groups
• Disciplines, Schools of thought
• Ecosystem Knowledge environment
• Unity of community, resources and activities

7
Entity

• (Information Object)
• Repository
• Collection of information objects
• Service
• Machine based activity on information objects
• Note everything is an entity!

8
Species

• This category is not applicable
• The only species participating is human kind
• IRs and aggregator services are non-living
  systems
• They do not reproduce, and do not evolve
• They are not dynamic (cf. 5.2.7.3)
• Alternative cultural artifacts

9
Interactions

• Types
• Machine-machine
• Machine-human
• Human-human
• Note activity might be the easier concept
• Direction of activity is not mandatory
• Definition of the nature of interaction is
  mandatory

10
Resource

• information object involved in an activity
• A service is an entity defining activities
  involving information objects
• A repository is a physical collection of
  information objects applying a specified set of
  services
• Any resource is an entity

11
Environmental factors

• Not applicable
• Interactions with internal entities (biotic
  factor) or external entities (abiotic factor)
  can be modeled as information objects or
  activities

12
Other

• Keystone, Niche
• Biodiversity
• n.a. because species are only humans
• Evolution
• n.a. because of missing reproduction
• Food webs and food chains
• n.a. because no digestion take place, information
  objects live on

13
Assumptions

• Specific types as encountered in DRIVER

14
Types of activities

• Service-service activities
• Service-repository activities
• Repository-repository activities
• Human-service activities
• Human-human activities

15
Types of users 1

• Researcher
• Generic information seeker
• Subject-based information seeker
• Repository provider
• Repository developer
• Service provider
• Service developer
• Repository platform developer

16
Types of users 2

• Librarian
• Publisher
• Advisor
• Advocat
• Funder

• DRIVER developer
• DRIVER data manager
• DRIVER supporter
• Infrastructure developer
• Infrastructure provider

17
Types of DRIVER services -- simple
UI Service
Recommendation Service
Community Service
AA Service
Enabling Layer
User Service
Search Service
Functionality Layer
Information Service
Index Service
OAI-Publisher Service
Collection Service
Repositories
Manager Service
Aggregator Service
Store Service
DRIVER Objects
Data Layer
18
Combinatorial Power Illustrated

• Researcher
• Generic information seeker
• Subject-based information seeker
• Repository provider
• Repository developer
• Service provider
• Service developer
• Repository platform developer

• DRIVER developer
• DRIVER data manager
• DRIVER supporter
• Infrastructure developer
• Infrastructure provider

• Librarian
• Publisher
• Advisor
• Advocat
• Funder

Combinatorial Explosion Occurs when a huge
number of possible combinations are created by
increasing the number of entities which can be
combined--forcing us to consider a constrained
set of possibilities when we consider related
problems. (Arbib) from Web Dictionary of
Cybernetics and Systems
19
Combinatorial Power Exemplified

• 10 users (human types)
• A minimalist number
• Disciplines not considered
• 10 services
• A minimalist number
• Number of repositories not considered
• Number of information objects not considered
• Reduction, of course, possible
• Adding constraits, logical grouping of services
  and users

c combinations n10 number of elements r10
redundant options, arbitrary c n! / r!(n-r)!
13168189440000
20
Interim result

• Complexity is a major challenge
• Ecology thesis 1 corroborated
• the inclusion of non-technical aspects in the
  frameworks (or models) for repository systems for
  the sake of easier comprehension

21
Sense of analogical reasoning

• Valuable if close correspondence between two
  models is given
• Close correspondence can be determined through
  detailed analysis
• Modelling is required for final assessment of
  correspondence

cf. http//en.wikipedia.org/wiki/Analogy or
also Gentner and Stevens, 1981
22
Ecological Modelling
conceptual model

• Ecology lends itself modelling approaches from
  systems theory (systems biology)
• ecological models would finally contain the same
  number of entities and causal relations as the
  repository system itself
• The simplification would be lost
• Systems theory itself is sufficient

formal model
If an ecological approach is embraced, it should
be pursued consistently. For example, ecology as
a explanantory model for cogntive processes
Gregory Bateson, Steps to an Ecology of Mind,
2000.
23
Alternative user-centric design

• Reduction of parameters
• Slave all parameters to specific scenarios (use
  cases)
• e.g. Harnad-Paradigm Provide a generic
  researcher free and ubiquitous access to a
  full-text of his choice
• Only one user type and one service type
• All other entities are organized with respect to
  this scenario

24
Alternative user-centric design
25
Practice requires exact models
OAI-PMH
Aggregator
search engine
26
e.g. UML specifications
Exact modelling of use-cases
27
DRIVER adopts modelling

• Combination of simple use-cases and generic
  architecture enables construction of exact models
• Specific solutions can be designed for specific
  problems

28
Interim result 2

• The ecology analogy points correctly to the
  necessity of modelling
• Ecological Modelling is itself is very formal and
  complex if taken seriously
• Ecological analogy may introduce even more
  complexity to the problems
• A system theoretic modelling is concluded to be
  sufficient in practice

29
Summary

• Repository ecology may be useful because
• It points out the necessity of simplification
• It points out the necessity of modelling
• Repository ecology may not be useful because
• it provides itself complex models that may even
  complicate understanding
• Conclusion
• Use ecology as metaphor but not as model
• Simplification through use-cases may be sufficient

30
DRIVER is not a pond
its a cultural artifact
Photo rom http//flickr.com/photos/hillsdalehouse/
21827297/ Small House in Pond of Fontainebleau
(Greg_e)