Jacsó’ s Cheers and Jeers for 2003

(See the abridged version in

  Information Today, January 2004)

Eugene Garfield

CSA-IDS

Ovid -  LinkSolver

PubMed

PMC and BMC

Emerald

CrossRef

DOI Foundation

e-psyche

PsycINFO

ALA Web site

This is an unabridged and richly illustrated version of my yearly  column. Click on the links in the text for annotated screenshot series about the cheered and jeered ones.

This time I dedicate the Cheers and Jeers column to digital products and services related to linking, and to the mother of all links: cited references in publications. Links are digital manifestations of the traditional intellectual acts of citing other works.

You certainly remember the scene from The Graduate. No, not that famous one with Mrs. Robinson, but the one where a family friend pulls away Dustin Hoffman at his graduation party, to give him a career advise, saying “I just want to say one word to you...just one word. […] Plastics”. Dear reader, even if you don’t read this column further, I just want to say one word to you, just one word. Linking.  Links build the Web.  Linking (citing) is an essential driving force in research, along with information derived from links (citations), such as the popularity (“citedness”) of  information sources. There are many to be cheered for treating links (citations) professionally, and there are some to be jeered for doing it in an appalling manner.

   Cheers to …

Eugene Garfield, the founder of  the Institute for Scientific Information (ISI), and the visionary of citation indexes for scholarly publications, who has relentlessly spread for 50 years the gospel of citation-based searching as an alternative/complementary option to keyword and descriptor-based searching through hundreds of his well cited publications and through  the family of citation databases which he launched and managed for decades.

His idea  is now being played out and mightily validated in the everyday experience of users on an ever widening scale. Savvy users have been  enjoying the backward and forward linking in the  well “Webified”  version of the ISI family of citation indexes (Web of Science), which made citation searching more accessible (cognitively, if not financially).

Now citation searching is offered through a number of other sources, including traditional abstracting/indexing databases (like PsycINFO and CINAHL), database collections of aggregators (like Ovid), digital archives of publishers (like ScienceDirect of Elsevier) and digital facilitators (like HighWire Press). These together offer more than hundred million actionable (hot-linked) citations. These links are the equivalents  of  the lianes of the rainforest. They  allow the users  to jump instantly from an article to its cited and citing sources and from there to the citing and cited sources of those, even across collections and disciplines – exactly as Eugene Garfield envisioned. See more about Eugene Garfield

CSA-IDS for the intelligent, innovative  and  comprehensive implementation of not only  links among cited and citing  references (backward and forward links in Web parlance) but also revealing and powerful citation counts in a number of traditional abstracting/indexing databases which  enhanced their records with cited references, as well as in the full text archives of the SAGE Collections in various disciplines. 

CSA-IDS displays in the results list of a query right above each title the number of times it was cited by other works. This is especially useful when scanning the result list trying to find the most promising articles. The “cited by” value  also appears next to many items in the list of cited references (along with links to abstracts within the same database or in another one of the CSA-IDS stable to which the customer subscribes).

This multiplies the effect of  a great idea which offers the chance for  making an educated choice in  picking the most cited (and thus presumably the most important) works from the result list as well as from the list of cited references.

These citation counts  could be made even more attractive if CSA-IDS would offer the option to sort the result of a query  in decreasing order of the absolute “cited by” values and even to calculate the adjusted “cited by” score per year. This score is the quotient of the number of citations received by the document and its age. . For example if an article published in 1999 has a “cited by” value of  160, and a book published in 1994 has a “cited by” value of 200, their respective adjusted score in 2004 would be 32, and 20, respectively (not counting the current year).

Similarly, an optional feature of sorting the cited references (if there are, say, more than 50 or 100 in a record)  in decreasing order of their “cited by” value, or adjusted score. This could bring yet another awesome  new feature to this outstanding software. I know that there are more pressing issues, but squeezing more juice out of an innovative feature is exciting. I played with this is idea in my atavistic way by  downloading results,  massaging them  into a spreadsheet, doing the above calculations and ranking the cited references by their score which I show in the linked file.   See more about CSA-IDS

Ovid for its LinkSolver software which offers utmost power and flexibility for systems librarians to develop scripts for creating links from an indexing-only database to abstracts in the many open access collections of publishers and facilitators (such as Ingenta and HighWire Press), to  articles in open access journals (such as D-Lib, Ariadne, Libres, Bulletin/Journal of MLA), subscription-based publisher archives, and practically any digital collection.

The software is exceptionally well thought out, smart, sophisticated, intuitive and immensely empowering and rewarding. I experienced it while putting on steroid an indexing-only database which had no DOI, SICI or other linking elements in the records, but with LinkSolver I could create generic query URLs with placeholder parameters, extracting appropriate bibliographic elements on the fly from the records, such as ISSN, volume, issue and starting page numbers, and add a hot-linked button to each record from the particular journal for launching a search for instant retrieval of the abstract or full-text of the source documents whose index entries appeared on the result list.   

I could formulate the scripts with many additional smart parameters (such as the start year since the journal or the abstracts are available in digital format). I generated links with little sweat (smarter cookies would break no sweat) to the free abstracts of JASIS, to the full text of several  journals of Emerald (such as Journal of Documentation, Online Information Review) which my library subscribes to, and to the best open access journals covered by the database. The extra beauty is that the link-generating scripts can be activated for the other databases to which the library subscribes. It was almost instant gratification to activate my links for example, in CINAHL, which has a surprisingly high number of records for information science and technology articles. This was my most rewarding creative experiment  in 2003 (amidst quite a few good ones). The software is not cheap, but it opens new dimensions of enhancing indexing and abstracting databases even if their producers don’t provide the links. See more about Linksolver

PubMed  for linking from within the MEDLINE records to the full-text of articles in open access journals (including the many top notch periodicals in PubMed Central [PMC]  and BioMed Central [BMC]) , making it the most gratifying among the many free and fee-based MEDLINE  implementations which also happily link you but not to the free articles but to document delivery services which even more happily charge a hefty shipping and handling fee and royalty even for many open access, instantly and legally downloadable  journal articles (like those from the British Medical Journal among the hundreds of others). See more about PubMed

PubMed Central (PMC) and BioMed Central (BMC) for digitizing entire runs of much respected journals, making them freely accessible by anyone and easily linkable as you can see in  the PDF and/or HTML archive of the Bulletin of the Medical Library Association (from 1911), or in the 1994-2000 issues  of the Journal of the American Medical Informatics Association, a top ranked journal in the information and computer systems category of the most current edition of the Journal Citation Reports. See more about PMC and BMC

Emerald for providing clear guidelines about linking into its archive through digital identifiers or by creating query URLs from bibliographic data elements. This made creating my links in LinkSolver even easier. Such practical aids will promote the use of links in bibliographies and manuscripts of any kind by any author or editor. See more about Emerald

CrossRef for offering free lookup of Digital Object Identifiers by anyone (although item by item not in batches), which again will promote the use of the most common and rapidly accepted link mechanism by authors, editors, as well as abstracting/indexing services. See more about CrossRef 

DOI Foundation for the excellent documentation and tutorials about the Digital Object Identifier, which is the key to reliable and standard linking. See more about the DOI Foundation

    Jeers to  

Database Access Group, LLC for presenting naive and absurd fiction as facts, about all aspects of the cheap e-psyche database. The tales told about e-psyche in direct and indirect PR materials of the company are so tall that they make the press conferences held by Baghdad Bob (the Iraqi Information Minister) look like high-fidelity, accurate and factual reporting of events  The president of the company, an industry veteran who had seen better times and had launched better projects kept making absurd promises and claims  (shown in my linked page)  in a letter to the editor and in a published interview by an unknown interviewer (who must have made Gary Condit pine for such docile person instead of Connie Chung). 

Claims that e-psyche has more than 4,000 journals were rampant. In reality it had less than a third of that - with sporadic coverage, yielding 40-50% records less than promised. As for the inclusion of cited references, customers,  online partners and the press were led to believe that “records in e-psyche are enhanced to include the references or citations as found in the original records” (emphasis mine). Perhaps in the fantasy of the managers (including another veteran) - if they refrained from looking at their database  Nothing could be further from the truth, even though this was the ultimate selling point of the database. 

In reality, much of the references are missing because first the references to pre-1970 materials were eliminated, then all the references to any books (books receive far the largest number of citations in psychology so this saved a lot of work, and emaciated the list of cited refeences). In a moment of desperation and insanity, the veterans limited the number of cited references to maximum 10 per article no matter how many items were cited by the articles. (In psychology, the average number of cited references in journals is about 33 in ISI. The number is higher, around, 44 in PsycINFO because it also indexes scholarly books and other monographs which have several hundred cited references ). Later even this limit was lowered and only a handful of cited references were added without any rhyme or reason as I illustrate in my linked pages below. 

The president of the company -when confronted about the massive differences between his claims and the reality- promised in May, 2002 that the e-psyche database would have 1 million citations and 150,000 records by March, 2003 and it will be updated by 6,000 records monthly.  As his other promises and those of chronic debtors.

By the end of 2003 (not by March)  there were barely more than 113,000 records, and e-psyche was  not updated since July, 2003. In 2002 only 45,244 records were added instead of 72,000 as promised. In 2003 less than 9,000 records were added. By the end of the year, there were only 1,565 records for articles  published in 2003. For PsycINFO this number was 50,579 before the last update in December, 2003.

The managers of e-psyche  kept cutting corners left and right soon after the initial release of the database. Now they should put this database out of its misery, then deep six it, and compensate their grossly shortchanged customers. It would be interesting to see if some of the victims of this database would form a nice class  and would test how limited really is the liability of this company. I would take the case but I am not licensed to practice, and I am not enough of  a veteran (yet?) to try to pull it off, but I would offer some assistance pro bono.   See more about e-psyche

The American Psychological Association (APA) for the incredible amount of syntactically wrong links to publisher and journal sites and for  the harried and sloppy implementation of the otherwise most welcome and laudable  inclusion of nearly 10 million cited references in 210,000+ PsycINFO records. Unfortunately,  some hundred thousands of the cited references are chopped off from the records (except from the OCLC implementation) because of the record size limitation of the MARC communications format which APA used to distribute PsycINFO  for 18 months. Most records start to conk out after 250-300 citations. Displaying a fraction of the citations in  most books and many articles must have mightily discombobulated users, and most librarians.  

Finally, in July 2003 APA started to deliver records in XML format which  has no size limits, but only OCLC was in a position to accept it. I have empathy for the apathy of the online services for not jumping hurriedly on the XML bandwagon after they have struggled for so long with the consequences of the extremely hasty (accelerated in APA parlance) and brutally ill-organized and sloppy launch of the enhancement project which could have earned my cheers had it not been done so carelessly.  

As I illustrate on my linked page, Dialog and Ovid may have been the most distressed by the trials and tribulations of the ever changing forms and tagging of the PsycINFO records. These may have made DIALOG  to add the descriptor subfield code 'd' to tens of thousands of descriptors in 2002 and 2003, causing a dmajor ddescriptor ddisorder, as well as  ddistress and ddepression among users, although I haven't heard or seen anyone  to complain about it anywhere. 

The incessant changes may have distressed Ovid so much that it added correction records without replacing the erroneous ones during updates, producing many duplicates. You will find nice examples in my linked page about this problem.        

It indicates the extent of the hasty implementation of the  enhancements that  PsycINFO indexers and or data entry operators have not heeded the warning in the APA style manual about the importance of checking the URLs, and supplied a mountain of syntactically wrong links which even technophobe users refrain from, such as using http:\\ instead of http:// prefix.  Not all  the external links have such slash errors, just 11,852 in the publisher field alone. This is complemented by a large number erroneous links for the same reason to journal Web sites.

APA should have focused  on the cited references project, instead of adding more records for 19th century articles, and upgrading records from the pathetic Mental Health Abstracts database which APA could not resist to acquire  (not accidentally without the usual press releases) in order to enrich  PsycINFO, which is akin to adding some salvaged furniture from a demolished YMCA hostel to a deluxe hotel. See more about PsycINFO

The American Library Association for launching the ill-conceived and ill-implemented new ALA site which resulted in the mass murder of  links in abstracting/indexing databases and in  bookmark collections of librarians. Special jeers are due for timing it for the Library Week festivities when more visitors than usual could see this embarrassment of the association and the profession. Although the site has seen some improvements since its deeply disappointing launch, ALA still has a very long way to go to demonstrate in its practice the state of the art technology written about  in some of its excellent journals which it  licenses to third parties, but (with sporadic exceptions) does not make available in a searchable, linkable digital format even to print subscribers, let alone to plain vanilla members.  See more about the ALA Web site

Back to eXTRA menu