SUBSCRIBE NOW!

Vol. 20 No. 5 — May 2003

Day-One Keynote

Each day kicked off with a keynote presentation. First up on day one was John Lervik, CEO and co-founder ofFast Search & Transfer (FAST). FAST builds Web search technologies that are used by some major companies, including AT&T, IBM, and Dell. In addition, FAST technologies drive the well-known search engine AlltheWeb as well as Elsevier's Scirus. In February, Overture Services, Inc. acquired FAST's Internet business unit for $70 million in cash plus performance-based cash-incentive payments for up to $30 million over 3 years.

Lervik's presentation described the current status of Web search engines and how they will develop and improve. Advances are needed in the four basic aspects of the search process: understanding the content, understanding the query, adaptive searching, and presentation of results. Lervik says that improvements in content analysis will come from language identification, better date and age determination, geographical coding, and "anchor texts" (hand-constructed index terms added to the metadata).

In order to better understand queries and so make better use of indexing and metadata, the query structure is broken down into the "head" (topic or concept) and the "container" (document type). There are of course millions of heads but relatively few containers, which Lervik illustrated with examples such as biography, image, article, etc.

Adaptive searching takes into account spelling variants and mistakes in both the query and the searched content as well as morphology, such as word endings and syntax. Semantics will aid in understanding the intent of the question.

The presentation of results will be improved by extracting and highlighting significant terms, such as people, places, and product names, and creating lists of them on the fly. Dynamic drill-down will deliver statistical analysis and information extraction from large results sets, instead of the poorly ranked lists that are currently provided.

Lervik's session was introduced by David Raitt, who is familiar to many in the information industry through his long service with the European Space Agency and his chairmanship of many conferences. He pointed out that the history of search engine development was littered with reinvention as technologies migrated from online to CD-ROM to the Internet. I couldn't help thinking that this keynote had described exactly the same issues of indexing, searching, recall, precision, and understanding users' queries that have been grappled with by librarians, A&I services, and information distributors since time immemorial. Onlythe language or terminology seems to have changed.

Day-Two Keynote

Richard Boulderstone, director of e-strategy for The British Library, presented the second keynote. Although the BL has collected electronic publications for some time, given the progress of legislation for the legal deposit of electronic material, his role is expected to increase dramatically over the next few years. (Seethis month's International Report on page 39.) The growth of this material, customer expectations, costs, competition, and government funding levels combine to create pressures on the institution. The BL's current administration is altering its vision from one of preserving an archive to one of fulfilling customer needs "to help people advance knowledge to enrich lives."

The variety of activities is certainly a challenge in itself. Aside from collecting and archiving currently published electronic materials, the BL plans to offer networked access to the collection across its six U.K. legal deposit libraries. Other content will be publicly available from anywhere. The collections span a vast range, including digitized historical or fragile materials such as the Gutenberg Bible project. Another project that is starting with a 100-site pilot is Domain UK, a Web archive for all U.K. sites. Yet another is "In Place," a collection of 100,000 images and sound recordings that illustrate the changing face of Britain and its people.

The BL's long-term goal is to have uniform storage and access strategies and technologies for this huge and varied collection. Whereas Lervik had stressed the importance of improving and increasing metadata, Boulderstone is looking to minimize the amount of metadata collected because of size and cost implications. It will be interesting to see how the two approaches compare in a year or two.

Day-Three Keynote

The third keynote described a totally different means of navigating the Web. Martin Dodge, of the Centre for Advanced Spatial Analysis at London's University College, studies the "geography" of the Internet. He believes that maps of cyberspace provide a useful and viable supplement to existing Web navigational tools. Dodge provided the audience with a tour of some existing and emerging examples while pointing out that no complete map is currently available today. He was anxious to stress that this was not a Hollywood-style visualization using 3-D virtual reality, but a more modern development of such classic information maps as Charles Joseph Minard's thematic map of Napoleon's march on Moscow or the London Underground route map. Maps of cyberspace are required to give better views of context, surroundings, hidden connections, and sometimes a bird's-eye view.

The key spatial properties of area, position, proximity, and scale—coupled with the graphical attributes of color, shape, labeling, etc.—are used to turn information into maps. An early example is the University of Arizona AI Lab's ET map (http://ai2.bpa.arizona.edu/ent) in which more than 110,000 entertainment-related sites are mapped as multilayered boxes of varying size with clustering of boxes of closely related content. While offering the caveat that all cybermaps are authored by someone or other and are thus subjective views of the world, Dodge provided several examples.

Visual Net from Antarctica Systems, Inc. has several visualization solutions, including Map Net (http://maps.map.net), which uses varying size areas to indicate the quantity of sites and information on various topics. [Editor's Note: For more on Antarctica, see the interview on page 1.]

SmartMoney.com (http://www.smartmoney.com/marketmap) presents a view of U.S. corporations by industry, size, latest stock price, and recent performance. The site uses area and color as well as click-through to current news items, earnings, financials, etc.

History Wired (http://historywired.si.edu), produced by SmartMoney.com for the Smithsonian Institution, is an experimental program through which one can take a virtual tour of 450 selected objects from the collections of the National Museum of American History. Many of the images and descriptions are of museum artifacts that are not currently on exhibit.

Netscan (http://netscan.research.microsoft.com) presents a map of Usenet discussions by topic with color-coding to indicate growing and declining subjects. Unfortunately, the Microsoft Research site was not responding at the time of this writing, so I was unable to try it out. [Editor's Note: The site is now accessible.]

Finally, network maps such as Kartoo (http://www.kartoo.com) and TouchGraph GoogleBrowser (http://www.touchgraph.com/TGGoogleBrowser.html) are intended to show links and interconnections between sites. Trying Kartoo for myself, I searched "INSPEC" in U.K. sites only, and, aside from understandable links to BIDS and EDINA, it directed me to the Ford Cortina Mk 2 Owners Club and e-Hernia.co.uk, a source for "frank information about hernias." Clearly, there's a way to go yet.

Dodge agrees that these tools are all early pilot projects, with none of them being his first choice for daily Web searching. However, they illustrate the potential for alternative views of Webspace. He's looking forward to the day when we have a London Tube Map-equivalent for the Web. Developments to look for include integration of mapping tools with the browser, maps of search engine results, automatic geocoding, multimedia searching, and reuse of information from earlier searchers (nicely described as "virtual footprints").

The general program provided parallel tracks with many practical sessions of tips, how-to's, and case studies. Topics included better searching, Web site design, teaching, information literacy, cultural development, authoring formats, resource management, and blogging.

The Exhibits

About 60 exhibitors showcased their products and services at the 2-day TLS event. Classified under technology, infrastructure, content, and services, exhibitors included publishers, security equipment suppliers, library automation services, subscription agents, and online service providers. Apart from all the usual suspects, I was on the lookout for exhibitors that I had not seen at other library shows. I came across a couple of interesting displays that were new to me.

SatWeb, Ltd. from Cornwall, England (http://www.satweb.co.uk), offers broadband satellite connectivity from fixed and mobile stations anywhere in Europe. Although useful for any organization that wishes to extend its corporate network to remote offices, SatWeb was targeting colleges and universities that wish to deliver greater outreach with their digital learning programs. They can provide mobile classrooms in locations that are too remote for conventional high-speed lines or to communities that cannot afford such services.

The concept is well-illustrated by a partnership called "Real" between Glasgow City Council, Glasgow colleges and universities, and various Scottish funding bodies. This project has brought broadband Internet access to deprived areas in Glasgow's large housing developments. "Real on the Road" is a mobile learning center that contains 10 PCs, printers, Webcams, learning materials, and satellite broadband Internet access. This 21st-century mobile library is bringing easy access, learning, and training classes to people who have never used computers or had such educational opportunities.

It's often said that although the medical profession is well-served with information resources, the busy general practitioner does not make the best use of them. To counter this, the U.K. National Health Service Information Authority has set up a National Electronic Library for Health (NeLH). NeLH (http://www.nelh.nhs.uk) contains more than 70 electronic resources, including MEDLINE, NHS guidelines, The Cochrane Library, research reviews and bulletins, and, new since last August, an online anatomy resource. Anatomy is a collection of 3-D pictures built up from body scans that can be rotated and their layers removed to expose underlying structures. The overall site is getting more than 2 million hits a month, with significant numbers from North America.

As I was leaving the halls, I mused on the taglines that most conferences and exhibitions seem to feel necessary to append to their name. ILI and TLS don't seem to use them, unless you count the line on the TLS registration form that says, "It's exciting. It's fresh. It's different." But that's a bit too much like toothpaste advertising for my liking. I much preferred a couple of examples that I noticed on the way out, advertising other meetings at the NEC. The first seemed relevant. "Making a Difference" sounds like a good mission statement for librarians until I saw that it was actually for the Prison Chaplaincy Conference. The second one reminded me of the pleasures and frustrations ofWeb searching and was used at an exhibition for one of Britain's most popular hobbies: It was simply titled "Go Fishing!"