100 likes | 202 Views
DESIRE 2. http://www.desire.org/. nicky.ferguson@bristol.ac.uk. The User’s View. Speed Repeatable … reliable Finding people, finding stuff Scalable view Explanatory technology Quality mark Community Centre. Resource Discovery.
E N D
DESIRE 2 http://www.desire.org/ nicky.ferguson@bristol.ac.uk
The User’s View • Speed • Repeatable … reliable • Finding people, finding stuff • Scalable view • Explanatory technology • Quality mark • Community Centre
Resource Discovery • Building an integrated infrastructure to help researchers find high quality information • Cross searching • different gateways, different protocols - WHOIS++, Z39.50... • Cross browsing • combining collections of several gateways • using ‘forward knowledge’ to dynamically create collections • Quality labelling information • machine readable descriptions of Internet resources • Putting it all together with RDF • Demo: rating, browse, search interfaces...
WWW Indexing and Harvesting Software • The Combine harvesting robot • distributed architecture • different components talk using client server technology • a modular design, easy to modify and extend • DESIRE 2 will improve and extend: • Metdata indexing - to include RDF • Range of document types • New summarizers and summarizer mechanisms. • Range of protocols harvested , to include NNTP and possibly FTP
Automatic classification • Report on state-of-the-art • Create a static database from a subject harvest • Test and evaluate different methods of creating the collection • Test and evaluate automatic classification methods • Simple matching with EI thesaurus and EI classification. • … with linguistic and heuristic improvements (project GERHARD). • Automatic classification using the Scorpion method (with OCLC). • Pilot service (alpha now running) demonstrating some of the above methods
Information Gateways Handbook and Workshop • User needs analysis/questionnaire and survey of current Information Gateway providers • Literature review • Collaborative activities and planning • Production plan, timetable and outline • Chapters now in production • Workshop for National Librarians - September
Directory Indexing • Goal: one distributed index for all directory protocols • Technology to be used the same as for the web indexing part of DESIRE II • LDAP crawler for LDAP and X.500 servers in NL resulting in a single index • served via the web (using AltaVista technology) and central LDAP server • moving from a central to a distributed index on a European scale
Web caching • Deployment: hands-on workshops & web site • Mesh autoconfiguration • Intercache communication • Testing of hardware & software
Directories News WEB Cross-search/browse Browser LDAP LDAP LDAP Crawler (harvester) TIO Server (query routing) Referrals Directory Client TIOs/CIP RDF, HTML... ROADS (subject gateways) Whois++ I/face not committed HTTP Gate- ways GILSSOIF COMBINE (harvester) ZEBRA(indexer) Z39.50 I/face NNTP DESIRE Components
Query interfaces Gateways Z39.50-Whois++ Index data Exchange (centroids, TIOs, GILS, SOIF, AltaVista, InfoSeek) Indexing tools (ROADS, Zebra) Vocabulary Tools Applic-ations Metadata Registry Tools Harvesting tools (ROADS, Combine, LDAP Crawler) Metadata extraction Gateways Whois++-Z39.50 Metadata storage/ management Metadata Exchange (IAFA, DC, RDF, Oracle...) Query interfaces Metadata Editors DESIRE Tools