Trip Report ECDL 2007
Trip report ECDL 2007 (European Conference on Digital Libraries)
Budapest, 16th sept to 21 sept 2007
Participants MPDL: Traugott, Ulla
Presentations etc: http://www.ecdl2007.org
Proceedings (1 paper copy) available at MPDL Muc, library
Overview on sessions[edit]
(see complete papers in proceedings)
- digital libraries and the web
- ontologies
- models
- multimedia&multilingual DLs
- grid and peer-to-peer
- preservation
- user interfaces
- document linking
- information retrieval
- personal information management
- new DL applications
- user studies
Tutorial[edit]
"Thesauri and ontologies in digital libraries" Dagobert Soergel, Univ of Maryland
unfortunately, quite dry and booklet-centered, no online examples
theoretical foundations, but no practical examples
booklet available in MPDL library, A-Strasse/munich
First part - structure and use
- Thesaurus functions
- why thesauri
- what is a thesaurus
- thesaurus functions
- thesaurus structure
- terminological structure: concept - term relationships
- conceptual structure: semantic analysis and facets, hierarchies
- Implementation, evaluation, resources
- Implementing thesaurus functions
- evaluation of thesauri (yahoo)
- thesaurus construction
- developing conceptual structure
- structure and processing of thesaurus data
some presentations ...[edit]
"digital preservation, archival science, and methodological foundations for digital libraries"[edit]
Keynote address by Seamus Ross, HATII, Univ Glasgow
imho, most impressing presentation (Complete keynote (paper) available in print at ulla's.)
digital preservation so far: research and problem solving paralyzed by socially created "risk clouds" => projects/research loose themselves in too broad areas of theoretical research questions, but no concrete solution for any problem found.
Issues are known:
- authenticity: apply diplomatic methods: quis? quid? quomodo? quibus auxilibus? cur? ubi? cuando?
authenticity/integrity: is duty of repository. reliability is duty of creator
- provenance: context of object!, pedigree/lineage, archival appraisal
Ross underlines:
- focus the research questions, tackle the open issues step by step, de-construct socially constructed clouds of fear on the "big black hole".
- try more experimental approaches in projects, intensify cooperation with engineers (grid)
- focus on actual risks, not perceived risks
[edit]
Univ of Bonn, Dept of Computer Science
query-by-lyrics retrieval system facilitates multimodal navigation in CD audio collections lyrics search engine, which allows lyric-audio alignment
"Automatic identification of music works through audio matching"[edit]
Nicola Orio, Univ of Padua
large music repositories (opera houses, radio channels etc) are challenging the exact identification of different performances of music scores.
=> methodology for music identification based on hidden Markov models
segmentation of audio recording
extraction of most relevant features of each segment
statistical analysis of the possible performances of a given score => specific performances are identified and tagged
"Roadmap for multi-lingual information access in the European Library"[edit]
Nicola Ferro, univ of Padua
How to design/model/implement multi-lingual information access functionality to TEL ?
Insight into current difficulties, architectural approaches and possible roadmap
"Opening Schroedinger's Library: Semi-automatic QA reduces uncertainty in object-transformation"[edit]
Lars Clausen, univ Library Arhus/Denmark
developed semi-automatic QA method for transforming objects for preservation issues=> method to evaluate possible shortcomings/Quality loss in transformation. several separate automatic checks of "aspects" of an object, combined with manual inspections => provide greater assurance that objects are transformed with little or no loss of quality.
"Texts, illustrations, and physical objects: the case of ancient shipbuilding treatises"[edit]
Richard Furuta, Texas A&M Univ
Nautical Archaeological Digital Library assist nautical archaeologists in reconstruction of ancient ships and study of ancient shipbuilding techniques. => Building virtual collections of references to fragments, timbers recovered from excavations, digitized ancient shipbuilding manuals (in various languages, with different semantics, with different focus of the manuals). Very interesting approach for relations between texts, illustrations and mapping to physical objects.
"Named entity identification and cyberinfrastructure"[edit]
Robert Kummer, Köln for Perseus Project/Tufts Univ.
Building on Perseus data (Standard TEI), build a VRE for the humanities, with focus on Greco-roman studies. how to integrate broad interdisciplinary and deep domain knowledge (classification schemes, authority lists) Customize general services (combine google map technology wiht expert knowledge on places/locations) find scalable methods for handling decentralized contributions.
=> all relying on underlying infrastructure: ability to canonically refer to people, places, objects in a text and to integrate knowledge on these entities from heterogeneous sources.
Using data structures CIDOC CRM, FRBRoo, MODS/MADS. (see paper!)
Workshops[edit]
"The interoperability challenge: Towards a European ecology"[edit]
Project based in UKOLN/eprints, aim is to develop an ecologically influenced approach to model non-technological entities and influences on building and networking repositories.
Presentation by Mahendra and John (Ukoln) on basic ecological concepts and terms used, and their mapping to a digital library environment. Three presentations by "external" projects, trying to map the ecological scenario to their current projects (DRIVER by Wolfram, PERX by Phil , RUBRIC by Caroline). Exercises and controversial discussions...
outcome: agreement to need for modeling non-technical aspects in the framework/models for repository systems,as complement to architectural approaches.
but ecological terminology not suitable, as hard to understand quickly and "overloaded". Simple definitions and proper visualization might be better approach. alternative models: ethnological models, sociological approaches (systems theory)
Overall Impressions[edit]
Conference was good possibility for aside talks
some interesting applications built on top of existing digital objects, but no answer to "how to create "good" digital objects?
research projects presented by very engaged young researchers, but sometimes the user needs were not visible
most presentations quite technical/mathematical (algorithm-centered;-)
no clear focus in overall session planning