Notes
Outline
Integration of Simultaneous Searching and Reference Linking Across Bibliographic Resources on the Web
William H. Mischo, Thomas G. Habing, and Timothy W. Cole
Grainger Engineering Library Information Center, University of Illinois at Urbana-Champaign
JCDL 2002
Overview
We have at our disposal the tools to create integrated digital libraries from the distributed digital resources environment in which we operate:
Standard retrieval environment (Web) and interface/client (Web Browser);
Standard transport mechanisms to connect heterogeneous content (HTTP, OAI, SOAP);
Standard metalanguage and tools for describing and transforming content and metadata (XML, DTDs & Schemas, XSLT, DC/DCQ, RDF, METS);
Standardized search/retrieval mechanisms (HTTP Post/Get, SQL, Z39.50, Object Oriented Databases);
Standard linking tools and infrastructure (DOI, OpenURL, CrossRef).
Candidate set of ‘best practices’ for IR.
Portals and Gateways
Role is to bring together and integrate disparate e-resources.
Provide a systematic ‘view’ of the information landscape, particularly full-text.
Two primary foci: robust search/navigation and the ability to link everywhere from anywhere in the environment of OPACs, A & I Services, full-text.
Central to this implementation is simultaneous search and reference linking technologies.
Distributed Information Environment
We live in a world of multiple, heterogeneous information repositories, resources, portals, and IR systems.
OPACs, A & I Services, Web Resources and Engines, Open and invisible Web, distributed heterogeneous publisher repositories, OAI resources, local data, multimedia data, GIS.
Need for document representation, retrieval, transmission, and linking middleware tools and standards.
Factor: changing landscape of Scholarly Communication and disintermediation of publishers and libraries.
Simultaneous search and reference linking as mechanism to integrate DL landscape.
Slide 5
Simultaneous Search Implementations
DialIndex from Dialog.
Ex Libris MetaLib service.
Endeavor EnCompass.
Innovative Interfaces MetaFind.
Ovid Multiple Search and reference De-Duping.
ISI Web of Knowledge.
Gale Corporation InfoTrac Total Access.
WebFeat.
California Digital Library SearchLight system.
Los Alamos FlashPoint system.
Fretwell-Downing partnering with ARL Portal and Monash.
Grainger Search Aid
Assist users in the selection of appropriate databases .
Normalize user search arguments and display search results from candidate databases.
Cross-database asynchronous concurrent searching.
Article level and e-journal Web site access to publisher full-text repositories.
Utilize OpenURL, CrossRef metadata database and DOI for reference linking at the article level.
Proxying of vendor systems and capability of ‘taking over’ the search in vendor native mode.
Slide 8
Slide 9
Slide 10
Slide 11
Reference Assistant Project
Utilize Search Aid simultaneous search and link capabilities.
Opportunity to explore interface and navigation issues.
Mimics the behavior of reference librarian.
Allows the application of ‘best match’ and ‘quorum searching’ algorithms.
Slide 13
Slide 14
Slide 15
Slide 16
Simultaneous Search Implementations
Shared Blackboard approach employing Independent Searchbots dedicated to searching information resources and passing results to Web clients.
Event-Driven, Asynchronous HTTP Queries from within a Single Script returning results to Web browser.
Shared Blackboard w/ Independent Searchbots
Searchbots are separate, independent processes, possibly across multiple servers.
Blackboard database allows the Web server and the searchbots to communicate.
Searchbots optimized for searching different resources. Can hold open connections and session Ids.
Very flexible, but complex.
Candidate for implementation as Web Service.
Event-Driven, Asynchronous Queries
Single, event-driven web server process, asynchronously querying multiple resources.
Uses WinHTTP from ASP and VBScript
Simpler, not as flexible. Search algorithms and processing coded in scripts.
This is the approach we currently use for our service.
Implementation of multi-step login and session variable passthru being investigated.
OpenURL-Based Services
Provides value-added links to the Ovid search results.
Uses CrossRef metadata database to look up DOIs.
Local OpenURL services:
Resolver.asp
Callno.asp
Journal.asp.
CiteParse.dll
An ActiveX DLL which can parse various Ovid citations and turn them into OpenURLs:
Tansu N. Chang YL. Takeuchi T. Bour DP. Corzine SW. Tan MRT. Mawst LJ. Temperature analysis … quantum-well lasers. [Article] IEEE Journal of Quantum Electronics. 38(6):640-651, 2002 Jun.
http://…/resolver.asp?genre=article&aulast=Tansu&auinit1=N&atitle=Temperature+analysis+…+quantum-well+lasers&title=IEEE+Journal+of+Quantum+Electronics&volume=38&issue=6&spage=640&epage=651&pages=640-651&date=2002-06
OpenURL and CrossRef
http://query.crossref.org/query/xref.cgi?pid=guest:guest&sid=null:null&report=NoneK&genre=article&aulast=Loke&auinit1=W&atitle=Epitaxial+growth+of+high+quality+AlGaAs/GaAs+structures+on+atomically+flat+GaAs+surfaces+in+deep+dielectric+windows+cleaned+using+atomic+hydrogen&title=Thin+Solid+Films&volume=402&issue=1-2&spage=43&epage=48&pages=43-48&date=2002-01-01
CrossRef metadata database elements: http://query.crossref.org/query/xref.cgi?pid=guest:guest&sid=null:null&report=NoneK&genre=article&title=Thin+Solid+Films&volume=402&spage=43
Conclusions
User reactions very positive.
The one-stop-shopping approach has been successful.
Users consider ability to link to full-text from citations in A & I Services and from references on publisher portals very helpful.
Technically, best approach appears to be a hybrid of asynchronous client interface with Web Services querying databases. Moves database middleware to Web Services and eliminates extensive custom script code for search and database query.