The Future of Narrative in Cyberspace
The opening address was given by Janet Murray and the title was "The Future of Narrative in Cyberspace". She has her own home page (http://web.mit.edu/jhmurray/www/) and is the author of "Hamlet on the Holodeck". She posed the question "Can there be great art in the new Media?" and she defined the properties of digital media as procedural, participatory, spatial and encyclopaedic. She suggested that the narrative of the future (in cyberspace) would utilise the techniques we now find in common use in electronic games such as no fixed ending, the ending is user defined; interactive; graphical; content continually under development; etc.
The unit of knowledge has shifted from being based on a number of words to the number of books, the global library and museum. We now expect to find everything at our fingertips and this highlights the problem of continually adding more data and yet the problems of navigation and organisation had not yet been solved! She suggested historical sites should not be interpretative but still needed to be organised (and accessible) e.g. journalistic sites present a problem - how do you view an article over time and the impact this has on organisation. These site present problems with stories with multiple perspectives including hyperlinks. She was presently completing a visual history, which would be organised by timeline rather than alternatives.
Janet asked the question How do we turn increased information into increased knowledge? Many fields have relevant views including: interface design, graphic design, literacy theory, instructional design, communication and library/information systems. She contented that the latter was best positioned to do the job.
She discussed two projects she was presently involved in: one for media studies and the other a site for Jewish women. These she contended required the Librarian to rethink cataloguing and access without existing frameworks. She said that in every project there is an archive, there is a dispersed resource and it should be designed for uniform intellectual access.
Janet has submitted a proposal to the Digital Libraries Group suggesting a research project which
investigates a single demon of intellectual access across TEI, EAD, Dublin Core, MARC and MPEG for biographical data to develop a controlled vocabulary for dates of birth, professions, milestone dates and other BIOREF which would be based on XML and would provide a coherent framework for retrieval and the display of biographical information.
CORC (Cooperative Online Resource Catalog)
Terry Noreault, Director of Research at OCLC gave a presentation on CORC (Cooperative Online Resource Catalog). He spoke about their developing approach to the problem of accessing WWW materials. He said that currently libraries in response to this problem do nothing, point to search engines, load records into the OPAC and build portal pages. There are a number of problems with portal pages. These include: extensive duplication of effort globally and within libraries (Yahoo indexes 1.2m sites, OCLC has 8300 members doing 100 sites each), the currency of domain names due to changing business focuses leads to broken links, changing content and new pages of course and the integration with local resources is difficult. (Still looking for my fully integrated user interface - Ed/kdb). The solution in his view was to construct a Web Approval Plan. This required structure, decisions about the type of resources to be indexed ( full text, index, abstract, portal pages). This requires a database of Web resources using high quality metadata and cooperative cataloguing. Their approach is to centrally harvest the metadata prepared for materials by participating sites using MARC/AACRS and Dublin Core i.e. the database record is a superset of these two standards. CORC is the tool used to build portal sites and twenty libraries are involved in this so far. He claimed 900,000 scholarly articles were published each year and asked the question how much is being spent on cataloguing these? 2,000,000 WWW sites are being added annually and he claimed it was within existing resources to collectively catalogue this material.
Supporting Users at a Distance
Marshall Clinton, Director of Information Technology at the University of Toronto spoke about "Supporting Users at a Distance". There are 55,000 students at the University of Toronto with only 8,000 on campus. It has the largest library in Canada with 8.5 volumes, 45,000 current serials, one central library and 40 college libraries. Twenty three percent of people are accessing library materials in the library, 5% from elsewhere on campus and 72% dial up. They are wanting to move from existing methods of support as they cannot use telephone support as the line is already used and email traditionally had a slow turnaround time. They have established a new service for remote users with Balisoft using 'Livesoft'. It was a small beginning with only the 8,000 proxy server users. The service requires a PC client to be loaded on the client workstation and allows clients to send messages requesting support to the support staff. They are moving to a Java client in 1999.
They are evaluating the types of questions being asked and the skills required to answer those questions. They are developing a knowledge base, and will be extending access to others over WWW. At present the support is provided by IT staff but they are looking to move this to the librarians mid 1999. With this change some staff are enthusiastic but some are insecure due to lack of skills. It's too early to tell what the client reactions might be to this change.
CIC Virtual Electronic Library
Beth Forrest from Warner University of Michigan and Barbara McFadden Allen who is the Director CIC Center for Library Initiatives spoke on the "CIC Virtual Electronic Library". This presentation outlined the government initiative in Ohio to allow library resources to be shared across the state. A very large project was underway providing linking authentication mechanisms between disparate library ILMS systems, Z39.50 searching and now moving to Z39.50 enabled document delivery services. The speakers spoke about the enormous logistics involved in delivering their project goals. The report is at http://www.cic.uiuc.edu/CIC/cli/z39-50report.htm. There are discrepancies between vendor documentation regarding Z39.50 implementations.
Internet Scout Project
Susan Calcari, the Project Manager for the Internet Scout Project at the University of Wisconsin, Madison spoke on "The ISAAC Network: The Internet's Selective Access to Authoritative Content". The Scout project has evolved from some earlier projects such as ELIB and Sausage (JISC funded). The goals of the project included building an infrastructure to allow for the sharing of metadata with the target being higher education. Indexing algorithms and assisted robots are being developed. The architecture is based on: Index - CIP(common indexing protocol) which was finalised in January 1998, Search - LDAP (lightweight access protocol) and
Interface - browser using forms. Content partners apply Dublin Core metadata (and need to keep this up to date!) and this is extracted weekly by a remote host. EDNA has expressed interest in participating and other collaborators cover USA, UK and Australia. Susan was dismissive of commercial competitors such as Yahoo as these are in the commercial domain with the metadata not being generated by Librarians. They are not using Z39.50 as everyone else is instead they are using the Library of Congress headings at present, but (considering) building a new classification system more relevant to WWW resources.
Citrix Winframe Terminal Server
Charles Dye from Indianna University/Purdue University in Indianapolis is an Information Systems Manager.
He discussed their implementation of Citrix Winframe Terminal Server and they have migrated 112 CD-Rom bays from CD-Net which is no longer supported. They have adopted a phased implementation - laptops in library buildings first, then campus staff, then dial in users and then they will be moving to full internet access (based on IP filtering). They do not use this to deliver desktop applications. The costs involved are $US1,000 per concurrent user, hardware for 25 users $18,000, citrix server $8,000, load balancing software $1,500, secure ICA $2,900. He suggests using as large a server as possible! They are moving to upgrade to WTS, metaframe and Windows NT with the cost not being an issue as campus agreement for NT workstation licenses exists. Using generic NT login may cause problems. They are averaging 7-10 concurrent users (after being live for 6 weeks) and he expects the service to grow to 100 concurrent users with 4 servers. The response times have been good but it is difficult to assess over modem links. It has had a big impact on plug in usage.
He wishes to extend to delivery of course materials to remote learners, part of architecture for delivery of on campus WWW based course materials and to support 140 instructional teams (librarians and others) to centrally mount CDs produced by academic staff as a front end to administrative systems.
Digital Library Scorecard
Brewster Kahle, the President of Alexa internet and the Chairman of Internet Archive spoke on the "Digital Library Scorecard". He gave a most energetic and lively presentation. He contends that the WWW now has a sufficient critical mass of data to allow computers to find and develop new relationships in data.
He defines a digital library as collection with selection, easy access, materials with organisation, preservation of the valuable and rare and an aid to the patrons. The WWW he contends has only to date addressed 1 out of 5 of these criteria - easy access!
In relation to collection with selection the volume of new material is a problem with it doubling every 8 months. You need 20m 'content areas' compared with 20m books in the Library of Congress. There is a low barrier to publishing for example the Starr Report where 1 in 7 www users had accessed the report within 5 hours of publication! The print version followed the WWW version (via CNN and not Yahoo!). However selection IS difficult. The second criteria that of access is well covered.
Regarding the third criteria materials with organisation he had a lot to say. Search engines return too many hits in result sets. Human cataloguing does not scale and search engines are developing into Media Malls adding features to 'keep you in the store'. Portals are indexing too little. He then outlined the Alexa system in which they are archiving all WWW pages (unless prevented) to a single repository, looking at the following questions such as who is behind the page, how popular is it, where is it, can you trust it? Alexa then provides via a browser add in toolbar access to services based on this information. Alexa uses a snapshot (taken by crawling) every 2 months but the metadata cannot be relied upon (e.g. modified date). Alexa uses two techniques to provide information via its toolbar. Usage paths (i.e. where did people who came to this site before go to next) and link structure - co-citation (who else cited this site, what else do they cite). From its archive Alexa can provide a copy of a version of a page missing with a 404 error. Concerns about copyright, privacy and legality arise in this context. It is totally (and unashamedly) supported by advertising revenues. http://www.alexa.com/
The first archive of the WWW was donated to LOC in November 1998 (12 terabytes).
In relation to the fourth criteria, that of preservation of the valuable and rare Brewster said that the average life of WWW content is 77 days before it no longer exists! He said the need for archiving of the WWW is not yet recognised and that Alexa are the only one with a copy of WWW resources for the 1999 US Presidential election - significant as the WWW was the primary delivery mechanism used. It is very important also in the context of the intranet. He thinks Alexa will in 5-10 years wish to change its focus to What's Changed rather than What's Out There.
In relation to the fifth criteria, that of aid to the patrons he asked How do we get Librarians back in the loop?! He said that usage paths analysis reveals that much time is wasted by patrons surfing the WWW and that Librarians are the one's who can give directions, build specialized collections and can be proactive in alerting users.
In conclusion he said the digital library does not address one, addressed two, three is starting to be worked on (going from key work searching), four is starting with models being developed and that the fifth criteria has not been addressed.
Kerrie Basman
Australian National University Library
March 1999
Send comments/suggestions/requests about this site ....