The U.S. Department of Energy (DOE) has signed an agreement with The British Library to partner on the development of a global science gateway that aims to make science information resources of many nations accessible via a single Internet portal, to be called Science.world. The signing took place in late January at The British Library, concurrent with the winter meeting of The International Council for Scientific and Technical Information (ICSTI; www.icsti.org). Raymond L. Orbach, under secretary for science of the DOE, signed the agreement with Lynne Brindley, chief executive of The British Library.
Recognizing the impact of international research efforts, Orbach stated: "It is time to make the science offerings of all nations searchable in one global gateway. Our goal is to speed up the sharing of knowledge on a global scale. As a result, we believe that science itself will speed up."
Brindley said: "We are delighted to be embarking on what we expect to be a very fruitful collaboration with the DOE to develop the Science.world resource. The British Library has a long history of delivering online information resources through international partnerships. …"
DOE's Office of Scientific and Technical Information (OSTI; www.osti.gov) will work with The British Library (www.bl.uk) and international counterparts to develop a prototype of Science.world in 2007. I chatted with OSTI director Walt Warnick about plans for the new gateway. He said they are working closely with ICSTI to spread the word to potential participants and would provide a status report at the summer meeting of ICSTI in France.
The new portal will use the model of Science.gov, which has served as the gateway to reliable information about science and technology from across federal government organizations since its launch in December 2002. The main advantage of Science.gov is that it lets users search for information by subject, rather than by the agency sponsoring it. The same will be true of Science.world.
Warnick said Science.world will provide a search of authoritative databases of nations, rather than a search of international portals. The planned resource would be available to scientists in all nations and to anyone interested in science.
Objectives of the Science.world initiative are the following:
- Search dispersed, electronic collections in various science disciplines
- Provide direct, seamless, and free searching of open source collections and portals
- Build upon existing and already-successful national models for searching
- Complement existing information collections and systems
- Raise the visibility and usage of individual sources of quality science information
Since its launch, Science.gov has continuously improved its search technology. In May 2004, Science.gov Version 2.0 introduced relevancy ranking with its QuickRank technology (see the NewsBreak at http://newsbreaks.infotoday.com/nbReader.asp?ArticleId=16445).
Science.gov 3.0 introduced MetaRank in 2005, which ranks results using custom algorithms applied to metadata or to the title, abstract, keywords, subject categories, etc. Science.gov 3.0 currently incorporates both QuickRank and MetaRank to return better search results.
In early 2007, Science.gov 4.0 will launch DeepRank, which will download and index the full text of documents to perform relevancy ranking for precision search. Grid computing will be integral, allowing distributed computers to communicate simultaneously and collaboratively. Warnick said these same ranking technologies will be used for Science.world.
Searches on Science.gov are powered by Distributed Explorit from Deep Web Technologies (DWT; www.deepwebtech.com), a small business located in Los Alamos, New Mexico. Abe Lederman, president of DWT, was present at the signing. A pioneer in federated search technology, he was excited about this initiative and said: "It's great to see this sophisticated search technology take on an international collaborative dimension. Leveraging the kinds of capabilities Deep Web Technologies designed into the U.S. Government's Science.gov federated search portal, launched in 2002, can help to provide a truly global resource for scientists and science-attentive officials, media professionals, entrepreneurs and citizens."
I wondered whether Science.world would be seen as complementary or competitive to Elsevier's science-specific search engine, Scirus. Joris van Rossum, head of Scirus, provided this comment: "Scirus, the most comprehensive science-specific search engine on the Internet, searches more than 300 million science-specific Web pages. A proven and well-established free search engine, Scirus searches across all science information resources including institutional repositories, journal sources, preprint servers, patents, author home pages, and other Web sources. Users from all nations can access Scirus for free via a single, easy-to-use interface. We welcome any effort to make scientific information easily accessible on the Web, however, from our perspective, Science.world will not provide users with any additional tools that Scirus does not already offer."But Warnick stressed the differences in the technologies underlying Scirus and Science.gov. The "crawl" technology of Scirus pinpoints scientific Web sites and Open Archives Initiative sources for search and then indexes the information. The deep Web searching in Science.gov does not rely upon a stored index built in advance but operates in real time, replicating the query and broadcasting it to multiple databases. He explained: "The deep Web search engine immediately reaches out to relevant databases at various sites, drilling down into these information centers all at once, organizing the info and returning results—in real time. So, not only are you plumbing databases, you're getting the most current results. As to content, the focus for the global science gateway will be on pinpointing and federating top government-sponsored research inside databases, rather than information gleaned from millions of Web pages."