Geen diatitel

1
? ? ?
From DOBES to CLARIN and beyond
Axel Horstmann Peter Wittenburg Erhard
Hinrichs VolkswagenFoundation MPI for
Psycholinguistics University of Tübingen
?
2
? ? ?
FACTS AND FIGURES

• Non-profit-making foundation established unter
  private law
• based in Hanover
• Not affiliated with the car manufacturer of the
  same name
• Founded by the Governments of the Federal
  Republic of
• Germany and the State of Lower Saxony in 1961
• Objective to support science and technology as
  well as the
• humanities and the social sciences in research
  and
• university teaching
• Assets about 2.45 billion euros
• Funding p.a. about 110 million euros
• One of the most potent private research funding
• foundations in Europe

3
? ? ?
FOCUS ON HUMANITIES AND SOCIAL SCIENCES

• Current funding initiatives
• (see KURZINFORMATION / BASIC INFORMATION)
• about 45 to 50 of the funds given to HSC
• Initiatives focussing on infrastructural support
  of HSC
• Kulturwissenschaftliche Dokumentation (closed)
• Archive als Fundus der Forschung (closed)
• DOBES Dokumentation bedrohter Sprachen
• Projects including infrastructural support of
  HSC
• Strategy building on digitization of endangered
  books
• Digitization of the so-called Aschebücher of
  the HAAB Weimar (in preparation)

4
? ? ?
"E-HUMANITIES" POSSIBILITIES AND PERSPECTIVES

• Strong interest in innovative approaches
• Funds available for projects involving
  activities towards
• "E-Humanities" (e.g. digitization of data,
  collections,
• archival material) within current funding
  initiatives
• Funding possibilities for meetings, workshops,
• conferences etc. focussing on "E-Humanities"
  (within
• the funding initiative Symposia and Summer
  Schools)
• New perspectives on "E-Humanities" (possibly)
  opened up
• within a new funding initiative aiming at
  Research in
• Museums (actually in planning) including to a
  certain
• extent digitization activities - and not to
  forget the
• Flagship "DOBES" ...

5
? ? ?
Concrete steps or Babylonian Tower

• we dont know exactly what eHumanities means
• we feel that mechanisms in research processes
  are changing
• rapidly with technological innovation as motor
• but we cant say we are now going to design
  eHumanities
• we probably can say lets plan further
  concrete projects
• and actions and see
• many excellent projects around let me just
  refer to the good sides
• of DOBES as one of these steps
• (Documentation of Endangered Languages funded
  by VolkswagenFoundation)

6
? ? ?
What is DOBES?
44 DOBES teams working fully distributed and
self-organized incl. linguists,
anthropologists, musicologists, ethno-biologists,
etc. In addition, VWF installed a central
archive Start in 2000
7
? ? ?
What changed in DOBES?

• handing over all data after a limited time to an
  archive was completely new
• and is an explicit step, although the results
  will not be ready
• there is a push to make data accessible to
  others from the beginning - also
• new for many and not without conflicts
• asking researchers to categorize and organize
  material according to
• agreed metadata was also new and still
  requires evangelization
• including multimedia in the documentation and
  dealing with audio/video as
• basis was kind of new and requires
  techno-knowledge

8
? ? ?
Which infrastructure by DOBES?

• a stable, reliable and open repository/archiving
  system handling 30 TB
• data storage not encapsulated and in open
  formats
• introduction of persistent identifiers to ensure
  investments in relating
• fragments
• a network of 12 centers worldwide included in
  data distribution
• of these 6 copies in centers with hardware
  migration strategy
• a number of web-based applications offering
  various ways to access the data

9
? ? ?
CLARIN/D-SPIN Challenges

• eResearch is about global collaboration in key
  areas of science and the next generation of
  infrastructure that will enable it (J. Taylor)
• goal is an open research infrastructure to
  overcome the huge
• fragmentation of language resources and tools
  and to offer them to
• research communities - in particular to
  humanities
• help tackling the LARGE challenges (multilingual
  societies)
• but also helping the individual researcher
• example align a transcription and an audio
  signal
• how many researchers know about how to do this
• see CLARIN/D-SPIN as a huge virtual marketplace
  of resources
• and tools that can be combined due to
  integration and
• interoperability solutions
• not forget Henry Thompsons (one of the XML
  fathers)
• don't have an agreed descriptive system in our
  domain

10
? ? ?
CLARIN/D-SPIN Research Infrastructure

• basis of big supermarket are classification and
• convincing organization principles
• based on 10 years of experience we know that
  only
• a flexible component model will be accepted
• seem to go towards a Federation of LRT producers
• that can make contracts with Identity
  Federations
• just one signature necessary to get all
  researchers
• with their home identity integrated
• have already setup a first small test
  federation (EC-DAM-LR)
• researchers dream virtual collection building
  and creating
• workflows flexibly - not trivial due to
  import/export aspects
• LREC showed that we know already a lot about
  the problem

11
? ? ?
CLARIN/D-SPIN Network of Service Centers

• need a network of strong and persistent
• centers of "new" type
• researchers will only adapt if they can rely
• on new mechanisms
• need to simplify the IPR/license situation

12
? ? ?
towards eHumanities

• CLARIN has gt 100 members from 32 countries
• in Germany 9 well-known centers and some more
  will join
• is an enormous challenge to make a real step
  ahead in CLARIN
• can we all together extend to eHumanities
  infrastructure or are we
• already close to collapse?

13
? ? ?
a few questions I

• will there be a separate infrastructure for each
  H discipline?
• NO
• there will be several shared services such as a
  PID registration and
• resolution service
• however
• building a joint infrastructure has to do with
  community building,
• trust, common language etc
• too big communities would not work
• so let's move on in TextGrid, DARIAH, CLARIN etc
• but let's have a close and fair contact to find
  synergies
• competition will become heavy and our
  competitors are the Googles
• of the world!

14
? ? ?
a few questions II

• will there be a single market place for the
  humanities?
• NO
• acceptance of a market place is dependent on
  classification and
• organization principles - as already said
• these are different in all disciplines
• so have to start from the disciplines in our
  solutions
• already difficult enough
• leave it to Semantic Web guys to enable
  cross-walk

15
? ? ?
a few questions III

• who will be the main players?
• of course the big libraries, archives and
  museums
• but what about the universities and big
  organizations such as MPG
• important
• we see new requirement profiles emerging
• kind of job sharing can be predicted
• of course close collaboration with innovative
  libraries such as
• SUB etc is required

highly specialized groups
highly specialized MPI departments
content centers
a number of domain MPIs
curation centers
MPDL few domain MPIs
computer centers
RZG, GWDG
16
? ? ?
a few questions IV

• key bricks for interoperability?
• we need open registries of all sort and smart
  registry frameworks
• schema registries
• concept registries (ISOcat - a creation of ISO
  TC37/SC4)
• relation registries
• etc
• however
• a very complex landscape seems to emerge
• how to make it usable by laymen?
• how to convince researchers to work with them?
• no one knows yet - we need to try out - what
  else?

17
? ? ?
Summary

• we need initiatives again and again to stepwise
  advance the borders
• it is now also time to transform existing
  knowledge into persistent
• infrastructures
• will need a lot of sensitivity and patience -
  RI building costs time
• emerging landscapes will have an underlying
  complexity
• need to offer discipline vocabulary
• need to hide complexity to a certain extent
• need to offer persistency
• Project solutions are not per se useful as
  infrastructure solutions!

18
? ? ?
End
in Germany we have already a good mixture with
TextGrid, DOBES, eAqua, DARIAH and
CLARIN/D-SPIN have to get together frequently
Thanks for the attention.