 |
 |
 |
 |
 |
 |
|
|
|
|
.
|
| . |
CNI FALL 1998 TASK FORCE MEETING
HANDOUT
|
|
. |
|
. |
|
University of Washington Digital Registry
|
|
Environment
- In 1998-99, we are migrating our database program from local loads and Z39.50 remote databases using our Willow clients to remote, web accessible databases using (mostly) vendors’ clients
- Looking for a way to integrate databases, ejournals, other electronica from many sources
- No web-accessible OPAC until March 1999
- Initial thought was to build an ejournals database to facilitate links to holdings and to fulltext. Soon expanded to encompass other electronic resources (indexing services, datasets, websites, locally networked CDs, etc.)
Design goals and parameters
- Design for existing workflows and reuse of existing data wherever possible. No budget for data entry.
- For phase 1, put everything through our Innovative Interfaces system and deal with single input stream into the Registry from III.
- Defined new local MARC fields for some non-standard data elements and purchased another MARC exporter from III. Already doing MARC export to feed our BRS catalog
- Use existing tools and development environment: NT Server, IIS, MS-SQL, Java, VBScript
- Multiple outputs: end-user searching, end-user customization, template-based HTML writers for automatic page generation, linking
- Plan to redesign sometime after web OPAC comes up 3/99
- Something has to be up for 9/98 release
Current Implementation
-
- Search/browse on common access points
- Add to My Gateway
-
- Available to users with University Libraries barcodes only (legal liability issue)
- Integrated with patron database
- Once authenticated, patron data is available to other request services on the site
- Mostly a roving bookmark/subscription list for now until more services are brought up
-
Page generation/HTML writing
- Most pages in the http://www.lib.washington.edu/subject/ hierarchy are automatically generated in whole or in part from database queries
- Only about 1/2 of subject areas had developed subject pages/pathfinders prior to release of the Registry. Links often appear on multiple pages, link maintenance problematic
- "What if" we could just generate subject pages automatically from the database and did link maintenance there
- Use LC classes and MARC country codes (assign additional if needed) and map to subject areas
- Provide easy way to override if subject specialist wanted to do more or thought generated pages weren’t adequate
- Can also publish arbitrary lists created with My Gateway using other templates (e.g., "core" lists)
- Generate "By type" lists, current awareness lists, etc.
Linking
- Links from A&I databases to holdings and to fulltext identified as features most desired in our last (1998) user survey
- Want to make it easy to move from A&I citation to document, whether that is local print copy, electronic copy, or copy obtained through ILL request
- Vendor implementations built around vertical integration (links to providers with whom vendor has a business arrangement), which restricts our ability to change vendors in response to consortial or market pricing opportunities, or in response to poor service
- Looking at implementing an intelligent proxy where user connects to database through our intermediate proxy URL that knows how to extract linking keys
- Proxy gets real page, parses it, inserts "Search for article" link
- If user selects "Search for article" link, proxy queries Digital Registry for holdings
- Proxy fetches article, holdings statement, or an ILL request option
- Issues
- Proxy needs to know about A&I page structure and about fulltext repository, so new A&I or fulltext sources need new components
- Changes in A&I or fulltext site design can break proxy
- Holdings maintenance: maintaining holdings statements (USMARC Level III required) for each electronic journal is relatively expensive for a single institution
- May use similar "page rewrite" proxy with authentication to support access by distance users without campus IP addresses
University of Washington Libraries, 11/98
William Jordan
bjordan@u.washington.edu
Steve Shadle
shadle@u.washington.edu
Alex Wade
awade@u.washington.edu
The original version of this document @ the University of Washington can be seen
here.
©
2008 by the
Coalition for Networked Information
ALL RIGHTS RESERVED.
webmgr@cni.org
| | |