Committee on Portals and Data Access Report

Committee Members: Leslie Johnston, Chair; Todd Burks, Jim Campbell, Joan Ruelle, Charlotte Scott

Charge for the committee

The charge for the Committee on Portals and Data Access is to strengthen the library's Web presence by exploring new models for providing a seamless on-line experience for as many library services as possible, e.g., electronic journals, digital library content, and the on-line catalog.  The group should

• ·identify existing and new functionality to be included in an on-line library
•  identify alternative access interfaces to traditional and digital library collections and services
•  set priorities for the development of on-line services
•  participate in the design of interfaces with the service developers

Considerations should include

• a review of the LofT Portal Team recommendations
• an assessment of how we can implement the features of SIRSI’s new Content Mamagement product(s) within our current infrastructure
• a review of the University's new portal infrastructure and how it might be useful
• methods for obtaining buy-in from both digital and traditional library specialists

Goals

Within this group, and meeting with other groups, such as the Virgo Steering Committee, consider the question of the definition of an online library for us and for our users, and the desirable services that it can provide.

Create a functional specification (not necessarily a technical specification) for a new "portal" onto the Library's collections across formats.

Review potential vendor products against the specification, and make recommendations that can be additional reviewed against the technical infrastructure.

Work with Digital Library Research and Development and Digital Services Integration to develop an interface to the Central Digital Repository.

Coordinate efforts at creating access to library services and/or collections additionally through the UVa UPortal.

Set priorities for implementation of new Library interfaces and UPortal implementation.

Summary Recommendations from the LofT Portal Committee

• Work with University Portal Team to develop a single UVA Portal.
• Develop a series of Scholarly Information Channels for the UVA Portal
• Continue development of the Library web site
• Maintain a "standard look" for all Library web pages
• Develop a series of new web-based functionalities to expand traditional services into an online environment.

One of the original goals for a library "portal" was a personalized Library content portal, where users could pick and choose between licensed resources and links to resources on the web to create a custom, personalized page.  When it was discovered in 2001 that ITC was beginning test implementation of a myUVa UPortal, the Library's programming efforts to meet this goal were halted.  The goal was instead restated that the Library would feed "channels" for services and individual content resources and web links into the UPortal to create a Library "page." 

The first step was cooperative development between the Library and ITC to create a generic library channel that includes a Virgo search box, a Library web site search box, and access to online reference pointers, which was launched in July 2002.  The feedback received through ITC during over the following six weeks was a strong request to add the ability to review one's user services record (books checked out, fines, request books, etc.), in addition to the already available discovery and reference tools.  Content channels were not even mentioned, save for channels for general reference works, such as the OED. 

In late August 2002 the above services were added to the general Library channel.  The UPortal has not been strongly marketed to the UVa community; and no feedback on use has been compiled or communicated at this time.  In late August a prototype was demoed for Library staff for a mechanism to send Virgo call numbers for selected results from a search to a cell phone, but at this point no additional content or service channel development is underway or definitively planned for the near future.

Given current Library resources and work that is moving forward to create a new point-of-access to the Library's collections, the group agreed that access to services via the UPortal is a much higher priority than access to content at this time.  It has been observed that other libraries that have developed personalizable portals where the focus was on the delivery of content over services that there have been very low use levels.  We are making the assumption that first focusing on creating a portal that provides centralized and personalizable access to services will gain higher levels of use that will only grow with the addition of content over time.

This is not to say that work in the content area should stop, as there are a number of potential projects to prepare both the VIRGO infrastructure and local digital content for more interaction with the UPortal system. 

• Work that is already underway to streamline vocabulary in VIRGO records that relates to computer files and Internet resources continue, as this will improve the ability to build simpler and more accurate queries
• The proposed work to transform the HTML in the VIRGO site into XHTML for improved interoperability with the UPortal and other XML systems should move forward.
• The proposed creation of an XML-based Internet reference site/subject area pages/Information Community directories be closely reviewed; as these reference/directory objects could potentially serve as channel content.
• Research should be undertaken to discover what mechanisms will be necessary to support communication between the Central Digital Repository and the UPortal to support the potential creation of content channels using complex objects stored in the Repository.
• The prototype for VIRGO to cell phone communication should be researched further, and the support for communication with wireless Infrared (IR) devices should also be researched.

The recommendation from the LofT committee that the Library web site development continue, and that a standard look-and-feel be developed are already underway, as guided by the Communications and Publication, Digital Library Research and Development, and Digital Services Integration departments along with various committees.

It is absolutely essential that the Library aim its development efforts toward a web-based discovery and delivery system that encompasses physical collections, digital collections, and licensed resources.  Given limited financial and personnel resources at the current time, the development of a Central Digital Repository and a single point-of-access to the digital collections is the first priority, followed by integrated access to digital and physical collections, and finally the addition of licensed electronic journals and databases.  It should be stated that priorities can shift, as the functionality of licensed resources and vendor systems develop.  It is expected that this will require a combination of in-house development and vendor solutions for discovery, delivery, and user tools.

It is key to stress that, while much discussion has been centered around a "single" web-based point-of-access to all collections regardless of format, that this concept is not in keeping with appropriate user services practice.  Rather, while there will be one mechanism for access to collections, a number of paths will also be available that point to that mechanism, including subject-based "lenses" on the collections provided through the branch libraries and subject librarians, and format-based access for focused research.

It is also key to stress the difference between this necessary service delivered through the Library's web site and the type of service that is provided is by the UPortal, which is support for the creation of personalized views of/interface to the Library's content and services.  The UPortal in no way replaces our development of services, only replacing any Library-based development of such a personalizable interface.  The systems employed in the implementation of the UPortal do impact our implementation of services that will require interoperability with the UPortal.

The Library is already facing competition for the portal and digital discovery "market" in the form of personal bibliographic management tools (End Note, Scholar's Aid, etc.) that can store and manage citations for physical and digital publications discovered online through Z39.50-enabled searching that can be configured in some situations to work with authentication systems to access licensed and generally available resources.  To retain these users we must be able to complete by providing a discovery system, a personalizable portal, and potentially supporting the use of these personal management tools with our systems.

This is a series of Library service priorities that should be supported and/or provided through the UPortal, the Library's general web presence, and integrated digital and physical collection access.

• Access to basic information about the libraries, including hours, location (including Grounds maps), and stack maps.
• Access to circulation information, both for items and for one's personal account.
• Check status of requests, and potentially handle LEO requests, which are currently processed by a separate Illiad system.
• Support, or, an a minimum, link to reference chat (currently moot given the current status of this service, but links to future implementations of this service are necessary).
• Sign up for tutorials (there already is a centralized web page, and requests are manually routed to the appropriate staff member).
• Access to Toolkit.
• Single point-of-access- searching of physical and digital collections together, including UVa, licensed, and selected web resources.
• Browseable directories of resources by discipline, including UVa, licensed, and general web resources.
• Access to lists of physical reserves (not electronic reserves) from VIRGO through Toolkit.
• Link to an improved introduction to the UVa Library for new students, staff, and faculty.

The are the recommended functional requirements for integrated collections access.  This is not a list of requirements that, if not met, will disqualify potential software solutions, but that must be reviewed as a series of priorities and trade-offs.

• Support for very granular control of authentication and rights management, either managed within the system or communicated from outside the system, such as the UVa proxy server or UVa-Anywhere.
• Must have automated indexing of new resources and a good local resource metadata maintenance interface, as well as, if available and affordable, a metadata maintenance service to provide updated links to the thousands of new and updated links from electronic journals every month.
• Facilitation of digital asset management and digital access management.
• Ability to create/record user profiles to control access to licensed resources.
• Full federated searching of all resources – Virgo, Central Digital Repository, Electronic journals and databases.
• Must de-dupe same results from searching multiple resources, something that is widely available in most vendor products.  An important point for the review of this functionality is the acceptable level of inexactness.
• Sorting of results sets by relevancy.  It is acknowledged that this is hard to implement, as it is not available from many vendors and there are interoperability issues – what relevancy criteria will work across so many resources and formats and metadata standards?  This is potentially a very complex mix of local data (the number of hits within a full-text resource, circulation statistics, and web user statistics) and ranking provided by certain licensed resources.
• Ability to create packaged, subset searches to provide to users by some criteria, such as by discipline or by format.
• Ability for users to create custom subset searches themselves on-the-fly that can persist as a part of a user's profile.
• The system administration functions must include load balancing.
• Vendor products must include data on scalability and load testing.
• Potential integration with personal bibliographic management tools.
• A Z39.50 front-end must be added to the Central Repository as that standard is already in place for VIRGO and many licensed resources
• Look at Open URL compliance for VIRGO, the Repository, licensed electronic journals and databases, and any potential vendor solutions.
• Consider indexing issues that we have in Virgo (such as indexing disparate metadata where ISSN and ISBN are not always present) and if they can be solved with new tools.

Recommendation - Software Implementation

Given the current budget situation, and the fact that the SIRSI portal product will not be available for review until mid-2003, the group is declining to make any recommendations about specific vendor product solutions at this time.