Reference Reviews

Péter's Digital Reference Shelf

June 2007

Title: dLIST
Publisher: School of Information Resources and Library Science, University of Arizona
URL: http://dlist.sir.arizona.edu/
Cost: free
Tested: continuously

The Context

Two years ago I chose dLIST as one of the picks in the Peter's Database Picks and Pans column . At that time, there were only about 400 documents in the depository, but its potential and importance for the practitioners, educators, and students of library and information science & technology justified in my eyes its selection as one of my picks.

DLIST was established in 2002, the same year as E-LIS, its only peer in the discipline, of which I just published a detailed review in the May issue of Digital Reference Shelf . (The Librarians' Digital Library deserves acknowledgment even if it seems to be limited to publications of Indian authors and a few papers by authors who presented their paper in India, such as the excellent presentation by Alma Swan in Bangalore about the potential of open access for science and scholarship in India. The OCLC Research Publications Repository is also an important collection of about 350 papers published by OCLC researchers. It is an institutional repository type, which in this case -by definition- is library and information science-oriented).

There are several other open access multidisciplinary digital repositories and depositories where preprint and reprint versions of papers in library and information science & technology pop up occasionally, but using a depository dedicated to this discipline can be obviously more convenient and expedient.

The Content

The last time that I not only used but also systematically reviewed this depository in the Spring 2005, it had less than 400 documents. In July 2007, there are close to 1,100 documents. This makes it still much smaller than the E-LIS depository, which has more than 6,000 documents, but still it shows the growth of dLIST. It must be mentioned that E-LIS has a very large number of non-English documents, and the size comparison should be made with the English language subset, but I could not determine it.

Although there are only 47 documents with the current publication year, documents of the previous year show an exceptionally high number of 281 deposited documents, a huge jump from the 165 items from 2005, which is the second most productive year in the collection, showing a steady increase from the 92, 98, 88, 58 and 54 documents from the previous five years. It is to be noted that about 60 items have no publication year. It may be surprising but there are a few items dating from the 1930s and 1950s – these are very well justified, as the oldest document in dLIST is the book of Five Laws by Ranganathan, which we educators often quote from, but could not expect students to read because few libraries had it. Now, there is no excuse.

More than 70% of the documents are formally published or in print papers. About 60% were published in refereed journals, and in the most influential of those such as JASIST, Information Processing and Management, Scientometrics, Library & Information Science Research.

The journal mix is impressive in dLIST. The other side of the issue is, of course, the language. All of the documents are in English, and dLIST should not apologize for it, even if I am the first to point out that many important papers in our profession are published in Spanish and Portuguese. These and other non-English-language documents are predominant in E-LIS, so the two depositories complement each other very well.

Small, but equally impressive is the collection of entire books and chapters of books. Many of them are current books, published in the past five to six years, and have some highly current gems such as Willinsky's book: The Access Principle: The Case for Open Access to Research and Scholarship published by the MIT Press in 2006. I wish some of his open access articles also were uploaded to dLIST, such as his very informative and substantial article about the indexing of scholarly journals published in 2001 in the open access Journal of Electronic Publishing (JEP).

Allow me a short detour here about how open access can be turned into closed access and no access by the whims of purely business-minded people. This article never received the acknowledgment and use it would have deserved — for a simple reason. Soon after its publication, Columbia University Press acquired the journal from University of Michigan Press. In its first move, CUP stopped providing access to the JEP archive and then there were no more moves and no more issues of JEP.

It must have been pure coincidence that Columbia University Press was about to offer a subscription-based service to the Columbia Guide to Digital Publishing, which was modestly described in a PR blurb as an "epic" work and the "first ever Bible of digital publishing." In my opinion, the seven volumes of JEP were much more useful about digital publishing than this subscription-based digital book, and I wish its excellent articles had been posted on dLIST in 2002. Willinsky's article would have deserved and needed open access the most exactly in that period, as papers typically get most of their citations in the second and third year after publication.

Luckily, Michigan University got back the rights and resumed publishing JEP from 2006 again as an open access, fully searchable journal after years of no access even to the back issues, including Willinsky's article. By the way, the "Bible of Digital Publishing" is not available online anymore. In print, you can find barely used copies for $6.08.

As for subject areas, dLIST, based on the analysis of the 90 subject terms assigned to documents (an average of 2.17 terms per document), the top most popular topics seem to be Digital Libraries (160), LIS Education (133), Scholarly Communication (117), Classification (73), Human Computer Interaction (69). At a closer look, some subject terms should be used in an OR relationship to get a more accurate popularity list by subjects.

For example, the terms Bibliometrics, Informetrics and Web metrics are assigned to 54, 19 and 15 documents, respectively. Some records have two of these terms so manual additions would not be fair, but a Boolean OR search ("ANY" in the parlance of its software) yields 76 records, getting ahead of Classification, but behind Internet (61) and World Wide Web (69) combined with the OR operator, which produces 113 documents. Any way you look at it, the topical coverage is broad enough and reflects well the hot issues of the current decade. 

The Software

dLISt uses the excellent, open source ePRINT software, as do so many other depositories and repositories. It offers a simple and an advanced search mode. The number of search options and filters in the advanced mode is unusually high. Beyond the usual choices of author, title, abstract, keywords, subject terms and publication year, it allows also the use as a search or limit criteria the conference title, the genre of the eprint (journal article, conference paper, books chapter), the publication status (published, in press, unpublished) and the peer review status (refereed vs not refereed) of the documents (where applicable).

Theoretically, the name of the department of the institution with which the author is affiliated is also searchable – but they are not easy to find as only a few records are enhanced with this metadata. There should be a note about this on the menu.

There are no options for term truncation and exact phrase searching, both of which make searching more efficient and relevant, such as in cases where you want to search for, say, library school but don't want items about school library issues, or the other way around in the results list.

There are browsing options by author, title, e-print type/genre, subject terms, conference title and publication year. I really would like to see browsing by journal title as well. In the browse list, you can see posting information, i.e. the number of items to which the listed subject term, author name, etc. were assigned.

The results can be sorted by author, title or publication year (either in descending and ascending order). The number of items per page cannot be controlled by the user (it is set to 10 per page), and there is no option for e-mailing select items from the result list which could be particularly useful when the user wants to alert someone else of the presence of some pertinent documents in the collection. This can be done on an item-by-item level from the detailed record, but not as a batch operation in one fell swoop for items selected. 

Overall the interface is intuitive (the layout of the advanced menu should be made smaller/tighter to avoid scrolling), the browse and search options are very good, the output options should be enhanced.

The usage statistics provided by eprints software are very useful. They give a good indication about the interest in the topic and the specific documents. They show how many times the abstract from the detailed record was looked up, and how many times the document was downloaded since it was posted. Breakdowns are available for specific time periods, and time periods distribution of users by countries. 

The only serious weakness is the lack of full-text searching (available in E-LIS, by the way), which would make the laudable idea of providing open access to important documents in a very focused, high-quality preprint and reprint collection even more appealing.

I hope that along with adding new documents at the intensity rate of 2006, some of the software enhancements also will be developed, especially the option of full-text searching – which is a decision by the implementors of the preprint/reprint collection, not a limitation of the underlying software.

— Péter Jacsó