One of my picks this issue is the SAGE journals archive hosted by HighWire Press. Its outstanding software and free abstracts represent the first significant social science subset amid HighWire's huge science and medicine journal archives. The other pick is the small but worthy DLIST collection of free full-text e-prints in library and information science. DLIST uses the eprints.org software, which made the posting of preprints and reprints of more than 600,000 papers--predominantly in physics, computer science, cognitive sciences, mathematics, and economics--feasible for researchers. The pan is LookSmart's FindArticles service. (This was already a pan after its harried debut, then a pick when it fixed the problems and offered open access to articles from several thousand professional journals and general interest magazines.) Once again its software shows signs of dementia.
SAGE ONLINE JOURNALS ON HIGHWIRE PRESS
SAGE [http://online.sagepub.com] is one of the largest publishers of scholarly journals in the social sciences. The majority of its journals joined the HighWire Press online search and delivery platform [http://highwire.stan ford.edu] in fall 2004, when the entire collection was free of charge until the end of October. The bibliographic citations and abstracts remain free, which is a big deal for those who cannot afford to subscribe to several social sciences abstracting/indexing databases. The family of SAGE Full-Text Collections is also available on the rejuvenated CSA Illumina system, but only for subscribers.
At launch there were approximately 57,000 full-text articles in the SAGE collection; by late March 2005, the number increased to nearly 73,000. (To put things into perspective: Although abstracts are also free through Google Scholar, it yields barely 10,000 hits when searching for SAGE articles.)
By my estimate, about 55 percent of the articles in HighWire's SAGE subset include abstracts. This is not surprising: Book reviews, news announcements, and letters to the editors have no abstracts. Neither do some of the substantial articles, especially in political science and sociology. For this reason, I pleaded with HighWire Press a year ago to offer a check box for limiting the search to articles with abstracts, but they were not convinced--yet.
The retrospective coverage varies. Journals in the psychology and education domains go back to the early 1970s; in the other disciplines, the coverage spans the past 20 to 25 years.
The SAGE articles cover not only the social sciences, such as psychology, sociology, political science, management, criminology, and education, but also communication, nursing and allied health, and urban planning--although to a lesser extent. Searching for the phrase glass ceiling yields nearly 300 articles from a variety of SAGE journals in sociology, management, psychology, and library and information science.
The software presents the results very attractively in a standard and a condensed format. The former makes it easy to scan the results list quickly and choose the most appropriate items based on the snippets from the full text. The options--to see articles from the HighWire-hosted journals, to be alerted when the article is cited or corrected in the future in those journals, or to launch a search for related items--are also free. Looking up cited references in the ISI databases is (understandably) available only if you have a subscription to the Web of Science database. The SAGE collection is yet another gem on the tiara of HighWire Press, which pioneered the role of digital facilitator and added a very important social science collection to its stable with lots of freebies for any users.
DLIST [http://dlist.sir.arizona. edu] stands for Digital Library for Information Science and Technology. Currently, it is more like a small digital shelf of open access articles, conference papers, and slide presentations about LIST topics. With that said, it is a promising initiative to demonstrate and spread the idea of self-archiving in our profession. DLIST is a service provided jointly by the University of Arizona's School of Information Resources and Library Science and its Health Sciences Library. Authors are encouraged to send their work for depositing in DLIST, which in turn uses the excellent and widely popular GNU EPrints self-archiving software, developed at the University of Southampton, to load and make the collection searchable. Although there were less than 400 documents in the repository at the end of March 2005, the year-by-year breakdown clearly indicates an encouraging trend: In the 1Q 2005, there were already more documents deposited than in the entire prior year.
You will find articles published in prominent sources, such as the Journal of the American Society for Information Science and Technology, Scientometrics, and Libri, as well as papers presented at the IFLA and ASIS conferences by well-known and novice researchers and practitioners. It is even more important that nontraditional materials are also deposited in DLIST. An example is the PowerPoint presentation made at the 2005 annual conference of the Association of Library and Information Science Educators (ALISE) in honor of one of the most influential people in our profession: Marcia J. Bates. You might wonder if a few PowerPoint slides are as worthy for inclusion in a repository as lengthy scholarly papers. My answer is a definite yes. To paraphrase the adage, a color slide can be worth more than a thousand printed words. Looking at Bates' citation profile through the prism of the Social Science Citation Index is the perfect image of homage by her peers, but it also whets your appetite to learn more about the excellent AuthorMap visualization software. DLIST may seem too young and too small to qualify for a pick, but I chose it because of its potential.
I panned FindArticles [www.find articles.com] a few years ago for making its parent company, LookSmart look dumb because of its incredibly low-IQ software. As an example, it searched for the pre-coordinated subject headings with a Boolean OR between the component words, so the term Convention facilities-Hawaii yielded 720,970 largely irrelevant hits because it ORed the concepts of convention facilities and Hawaii. After the not-so-brainy developers were let go and management changed, the software was much improved and was on par with the valuable free full-text content licensed from the Gale Group. A year or so later, I chose FindArticles for a pick and praised it for the abundance of free and good quality content. In the latest incarnation, the free collection is much reduced, but still it has substantial backfiles of worthy journals, such as Library Trends. Added fee-based content comes from Gale's Goliath database, HighBeam Research, and KeepMedia. The software, however, has relapsed to dementia, seriously hindering access.
The Advanced Search Template once again shows lethal signs of ignorance and mightily discombobulates the user. For a starter, the year delimiter begins at 1950 even though there is not a single article before 1973. There are serious and voluminous problems. When searching for articles with the word database anywhere in the documents but restricted to, say, ON-LINE as the journal (selected from the journal list), the software comes up with no result, even though I clicked the "Show premium content" radio button. It turns out that the software limits all searches initiated from the advanced query form, no matter what you specify.
Searching by author name Ojala (again asking for premium content in addition to free), the search is run in the free subset and finds one such article by Tommi Ojala. Once I unchecked the free articles check box on the top of the result list (which should not have been checked, but the software did not recognize the setting on its advanced template), the software got a lucid moment and found some articles. Remember, this works only if there is at least one matching free article for your query so that you can uncheck the checkbox on the result list, which luckily is understood by the software. If there is no free article by the author, from the journal, or on the topic, you are at a dead end. You have four limit options to search: in the title, in the body, in the title or body, and anywhere. What's the difference between the third and fourth option? Nothing. Maybe the third option could be used to limit the query to the abstract.
The software proudly offers an option to limit your search to articles that are longer than 0, 1, 2, ... 10 pages. (In computer science you start counting with zero, but in this context it is safe to assume that for this value a > 0 attribute can be taken for granted.) Then again, the software seems to consider all (or almost all of) the articles to be one page in length. If you use this limit to search on any subject for lengthier articles--say, longer than five pages--you are likely to have 0 results. Why? Because according to the software, there are only 1,888 free and 99 premium articles out of the allegedly 5 million articles that meet that criteria. Someone should have tested this pathetic search engine before unleashing it onto the public.
As for the tag line "Find 5,000,000 articles not found on any other search engines," it is as intelligent as the software and as true as a used-car salesman's claim. Most of these articles are available on the search engine called InfoTrac OneFile, a Thomson Gale product, with the others showing up on HighBeam and KeepMedia. Keep in mind, too, that many articles are available through other general interest databases that are free to thousands of public, school, college, and special libraries' patrons. Many are also available directly from the publishers.
Scanning the journal list, you spot many titles that have been free for a much longer time than what is covered in FindArticles.com, such as the British Medical Journal, IBM Systems Journal, and IBM Journal of Research and Development. (These last two have 45- and 50-year free coverage, offered by IBM in PDF.) The articles are all several pages in length despite the fact that FindArticles.com claims them all to be one page long.
As for the number of the articles in the tag line, I just don't believe it, but never mind. As long as the software makes sure that you can't find them or can't find them easily, it is a moot question. This software will do in the worthy full text content still left in the free domain of FindArticles.com.
University of Hawaii
Peter Jacso [firstname.lastname@example.org] is professor of library and information science at the University of Hawaii's Department of Information and Computer Sciences.
Comments? E-mail letters to the editor to email@example.com.…