230 likes | 383 Views
OpenURL: Pointing a Loaded Resolver. Kristina Long Simon Fraser University Access 2003. Outline. Projects OpenURL 101: Version 0.1 Version 1.0 Link Resolvers: knowledgebase link resolving logic administrative interface. OpenURL and Link Server Projects. GODOT
E N D
OpenURL:Pointing a Loaded Resolver Kristina Long Simon Fraser University Access 2003
Outline • Projects • OpenURL 101: • Version 0.1 • Version 1.0 • Link Resolvers: • knowledgebase • link resolving logic • administrative interface
OpenURL and Link Server Projects • GODOT • Holdings, requesting and full text linking software used by COPPUL and ELN • Parses incoming OpenURLs • GODOTjake • Full text link server used by GODOT and written by Todd Holbrook • CUFTS • Completely redesigned link server written by Todd Holbrook • Parses incoming OpenURLs. Provides HTML output for links directly from citation or XML output for use by software like GODOT.
OpenURL version 0.1 • Syntax for a URL containing bibliographic information • Version 0.1 is simple to understand and implement • 28 data elements for describing seven different genre: • Bundles: journal, book, conference • Individual items: article, preprint, proceeding, book item • Includes field for local information (‘pid’)
OpenURL version 0.1 • Example (without escape encoding): http://godot.lib.sfu.ca:7331/godot/hold_tab.cgi ?sid=AMS:MathSciNet &genre=article &title=Phi Delta Kappan &atitle=A tale not widely told &volume=74 &date=1992 &pages=196-7 &issn=0031-7217 &pid=<an>743928749</an><authors>Doe, John; Brown Peter</authors>
OpenURL Version 1.0 • National Information Standards Organization (NISO) trial of 1.0 in progress - trial ends November 2003 • Framework has been generalized so it will be of use for more than bibliographic citations (eg. legal documents, genome sequences)
OpenURL version 1.0 • new genres (eg. Journal issue, Dissertation, Patent) • request can be sent ‘by reference’ • extensible • richer metadata options: • Multiple formats: Key-value pairs, XML, MARC • Element sets for different types of items
OpenURL version 1.0:Metadata Formats Core:Metadata Formats ori:fmt:kev:mtx:bookKey/Encoded-Value Metadata Format For Books ori:fmt:kev:mtx:dissertationKey/Encoded-Value Metadata Format For Dissertations ori:fmt:kev:mtx:journalKey/Encoded-Value Metatdata Format For Journals ori:fmt:kev:mtx:patentKey/Encoded-Value Metadata Format For Patents ori:fmt:xml:xsd:bookXML Metadata Format for Books ori:fmt:xml:xsd:dissertationXML Metadata Format for Dissertations ori:fmt:xml:xsd:journalXML Metadata Format for Journals ori:fmt:xml:xsd:MARC21Library of Congress MARC XML Metadata Format ori:fmt:xml:xsd:oai_dcOpen Archives Initiative Unqualified Dublin Core ori:fmt:xml:xsd:patentXML Metadata Format for Patents
OpenURL version 1.0:Matrix defining the KEV Format to represent a dissertation
OpenURL version 1.0:By-Value http://resolver.mylib.ca/resolver ?url_ver=z39.88-2003 &url_ctx_fmt=ori:fmt:kev:mtx:ctx &url_val_fmt=ori:fmt:kev:mtx:journal &rfr_id=ori:rfr:myid.com:mydb &rft.genre=article &rft.title=Phi Delta Kappan &rft.atitle=A tale not widely told &rft.volume=74 &rft.date=1992 &rft.pages=196-7 &rft.issn=0031-7217
OpenURL version 1.0:by Reference http://resolver.mylib.ca/resolver ?url_ver=z39.88-2003 &url_ctx_fmt=ori:fmt:kev:mtx:ctx &url_ctx_ref=uri:http://www.example.org/objects/1234.txt
Implementing Incoming Version 0.1 OpenURL Links • Minimal implementations (eg. ISSN only) • Need parser for local information in ‘pid’ field • Quality of data in OpenURL varies among vendors • Vendor documentation often minimal or non-existent • Vendor OpenURL linking feature can’t be customized (eg. link image or wording, window size, window name)
Link Server Software • Most common application of OpenURL • Components: • Knowledgebase • Link resolving logic • Administrative interface
Knowlegebase • Essentially a title list database: • title • ISSN • e-ISSN • full text coverage (date, volume and issue) • embargo period (days or months) • vendor specific ID/code • base URL • DOI • Local configuration information may also be considered to be part of knowledgebase
Knowledgebase: Vendor Title List Problems • Lack all the required fields • Poor quality or only intended for humans to read (eg. coverage dates in inconsistent formats) • No easily machine readable format provided (eg. tab-delimited) • Not updated frequently • Not easy to find on website or only available by email request
Link resolving: Storing Linking Information • How to store linking information? • Code (eg. Perl) using provided utility functions • Link templates • Need a way for libraries that want to add their own links to new resources to do so
Link resolving logic: Issues • Do you pre-fetch the link? • Pro: minimizes bad links • Cons: • extra logic for linking module • if not a proper API, then wording changes on vendor’s end may break linking module logic • increases vendor/publisher site hits • increases time to return results
Link resolving logic: Issues • Do ISSN only lookup or do title as well? • If doing title matching then: • do you use a title authority file? – may be time consuming to maintain • fuzzy matching vs. false positives
Link Resolving Logic: Issues • How to handle multiple hits from the same provider (eg. article appears in multiple Ebscohost databases)? • Need provider level de-duping.
Administrative Interface: Simplifying Configuration • Minimize the amount of information required • Keep screen specific to resource • Make details of the implementation of the links transparent • Provide site groupings for multiple branches or for consortia
Administrative Interface: Data Merging Issues • Which title list details need to be able to be overridden at the library level (eg. coverage, journal title)? • Should the data displayed for a screen be the merged, the local or the global data? • Consider how to clearly display the merged data • SQL not good at handling overriding data, so must be done by application code
Administrative Interface: Minimizing Duplication of Work • Especially an issue with ‘partial collections’, eg. Springer, Highwire • Allow integration with external subscription databases (eg. Custom e-journals databases, Serials Solutions) • How to implement integration? • Live lookup on remote database. • Upload dump of data (manual or automated) • Either way must maintain mapping of resource names
References • OpenURL Syntax Description(http://www.sfxit.com/openurl/openurl.html) • Implementing OpenURL 1.0, by Oliver Pesch (http://library.caltech.edu/openurl/Presentations.htm)