The second Virtual Reference Desk (VRD) conference attracted many more delegates than last year, and more than anticipated by the organisers. Last year around 250 attended, this year over 500 attended. It provided a very valuable opportunity for librarians, "ask-a" services and IT developers to exchange ideas and discuss key issues in the delivery of online information services. Four major areas were discussed -- nature of services, information infrastructure required to operate a services, software and service quality/standards. Many issues were very controversial -- the most argued being service quality. The two ends of this spectrum were represented by Ask Jeeves, which was widely criticised for answering questions from an automated "knowledge base" and the individual human answers to AskERIC. Pennie Finney, from Ask Jeeves, reminded the conference that any service that answers 4.5 million questions per day, as Ask Jeeves does, cannot provide this service based on full human answering systems.
What is virtual reference? At the conference, presenters and delegates reported on three different types of virtual reference services, and it was often difficult to determine which sort of virtual reference service was being discussed.
The types of services are:
- E-mail reference services. E-mail services are the longest established, and with the National Library of Australia’s service celebrating its 6th birthday next year, it is one of the longest provided service. Based very much on moving traditional written enquiries to a new environment, these services provide answers to the organisation's designated client group using traditional reference library skills. Key developments in this area in public libraries and the "ask-a" services included the move to a more call centre type approach along with a very quick turn around on questions (aiming for 5 minutes), short responses and referral to electronic resources. This is the basis for most collaborative projects such as the Library of Congress initiative Collaborative Digital Reference Service (CDRS).
- Online/chat to the librarian reference requests. Online/chat is still new and is generally still in an experimental stage. Chat room technology and telnet have not worked well thus far, both because of technical limitations and because of the difficulty of managing easy access to a bank of specialist, reference librarians. Some of the online/chat projects have found that the service has generated only low usage by clients; often trials that lasted a month resulted in only a handful of questions being asked via online/chat. Research from Temple University suggests online/chat only works well for short questions and as an initial access point. Software is very slowly improving, with that of the LSSI (Library Sysntems & Services http://www.lssi.com/virtual/) the only bright light on the horizon. This software is being implemented in San Francisco Bay libraries, which service a huge population, but the service is not yet provided on a 24 hour basis. The software enables routing of questions online in real time by subject strength. Issues of user demand, service models, business models and service quality are at a very early stage of understanding.
- Fully automated solutions that provide access to an answer base/knowledge base. Best represented by Ask-Jeeves, these services provide access to a storehouse of information and will do a keyword match and send information to requests. These automated solutions scale well; however issues of quality and relevance of returned information still require significant work. The motivation for the development of these services is different from the above two service categories. While most email reference and online reference (asynchronous and synchronous) services have emerged from traditional libraries or scholarly/subject communities seeking to enable a broader group to access their information resources, more complex business opportunities are spurring the development of fully (or mostly) automated reference services. The busines groups are the only groups really looking seriously at issues of scalability.
- Reference services standards. Just as standardising or measuring the quality of traditional reference services in libraries has been impossible to accomplish, service standards and quality assessment are proving very difficult in digital reference. Minimum standards in term of response time and user satisfaction were applied in a number of cases, but do not give a full picture. In other studies, word count was used to perhaps indicate depth, but as Miwa’s paper pointed out, users have multiple goals and different situational variables (including technical skills) so quantitative assessment will only give part of the picture. An excellent paper was presented by Jones summarising the research into virtual reference services -- library, commercial and organisational/subject community.
- Software. As indicated above we are still working with very early versions of software. The Internet Public Library, LSSI Knowledge Bit and Answerbase presentations all confirmed that standards need to be further developed for questions and answers if data is to be shared or automation used to its full potential.
- Who pays? This very important question underpinned many of the panel discussions. For organisations who have a mandate to provide public information, the demand may be huge, as may retraining staff and changing services to meet the needs of all users. For other services, options for funding, both government and private, are critical. AgNIC, a collabroative project of the National Agricultural Library, land grant universities and agricultural organisations (http://www.agnic.org/) is a good example of a service provided to a defined audience (agriculture workers and researchers) that has moved to a broader basis with an international audience.
- Partnerships. Perhaps the strongest message to take home from the conference was that librarians can’t do everything by themselves. Collaboration -- for subject strengths, to move towards 24x7 service, and to seek funding -- was needed to move forward.
The conference offered a very eclectic mix of technical, operational and policy issues. The conference papers will be available shortly on <http://email@example.com/>. The next conference will be in Atlanta in 2001.
Click here to return to the D-Lib Magazine "In Brief" column.