Dr. Strickland has hit the nail on the head: NIH researchers should be making more effective use of bibliographic-management programs and computerized reference-update services. I have a somewhat different perspective on the paths to getting there, but it doesn't alter his excellent advice.
I prefer updating my references via NIHnet and Gopher because the service is free and requires no maintenance on my part. There are some inconveniences to using Gopher, especially in having to restate searches every time and getting references one at a time. On the other hand, for a quick and easy search of the literature, it's usually quite adequate. Gopher also allows you to search references as far back as eight weeks. This is not the case for reference services provided via software installed on individual or local-area network (LAN) servers, which only allow you to search a single week's data at a time. Furthermore, in contrast to Strickland's experience, I've found that references obtained from NIH Gopher are incorporated into database files just as easily as those from Grateful Med.
Researchers who routinely browse the literature using the same complex search schemes can make use of Auto-Gopher, which will mail you the results of automatic, complex searches on a regular schedule that you determine -- weekly or monthly, for example. In addition, this service retrieves your references in a single file that can be taken up by any bibliographic-management program that can read Medline format. In addition to the automatic search service, Auto-Gopher also provides special, one-time searches for researchers who want to get a particular set of references in a single file and not have to retrieve them one by one as they do in a Gopher search. The only drawback to Auto-Gopher is that it only runs via the Helix computer. That means it's not available if you use method other than the Helix computer to access Gopher files, such as a browser program.
A disadvantage to buying Reference Update or Current Contents software for an individual computer -- besides the cost to the lab -- is that it takes up disk space and also requires maintenance (someone has to copy the updated data as they arrive). As for access via a LAN, the major shortcoming of Reference Update is cost. In fact, just three LAN licenses, each accommodating 10 users, would exceed the cost of supplying all of NIH with Reference Update over Gopher. Furthermore, just as for individual computers, installing reference-services updates on a LAN server takes up hard-drive space and requires maintenance. Not every group has LAN coordinators to do this.
Nonetheless, using reference service software such as Reference Update or Current Contents on your individual computer or via your LAN or Internet is more sophisticated than anything you could hope to do on any Gopher. So, if you have requirements that can only be met by this software, it may be worth the expense. (Note: I think that Current Contents, although available on disk, is not available by Internet.)
Current Contents' major disadvantage is that it displays references with the authors' names, titles, and journal names all in upper-case letters. This results in the references being taken up in all-upper-case letters by bibliographic-management programs such as Reference Manager or Bookends Pro -- a style not accepted by most journals, which usually prefer an upper-case/lower-case format. To change the style while in Current Contents, you have to edit the individual references and change the entries back to lower case. To save typing, you can open the original reference file(s) with a word processor that will "de-capitalize" everything past the first character. Then, you re-save the file as text and take it up with your bibliographic-management program. You will probably still have to do some editing, but it will be much less than if you made all the changes from within your bibliographic-management program. Also, mistakes in re-entering data will be avoided. Only the End Note bibliographic-management program strips the caps after the first letter, and even that program sometimes makes mistakes on unusual words.
The biggest shortcoming of Reference Update, no matter how you gain access, is that far fewer journals are referenced by this database than by Current Contents. However, depending on your field of interest, this may not be significant, and there may also be journals in Reference Update that are not represented in the Current Contents database. Try an experiment: do the same search for the same week on both Reference Update or Current Contents. How many references did you get, comparing them to each other? If Reference Update seems to be doing a good job, then stick with it because its reference format is far superior.
Recently, the company that owns Current Contents, The Institute for Scientific Information in Philadelphia, bought Research Information Systems of Carlsbad., Calif., which produces Reference Manager and Reference Update. So, there's hope that the best features of both reference-update programs may soon be combined into one product.
There are advantages and disadvantages to every bibliographic-management program, and Reference Manager is no exception. Although very powerful, Reference Manager is expensive, slow, and more difficult to use than some other programs. You are also limited to a single "registered" database per copy of the program, which could either be an advantage if you want all lab-related references in a single location or a problem if your research group requires multiple reference databases. For example, in my work with the BioInformatics and Molecular Analysis Section of DCRT's Distributed Systems Branch, I like to create a database for each paper I am writing by taking references stored in previous databases. No other program that I am aware of besides Reference Manager limits you to a single database, and some, such as EndNote Plus and Bookends Pro, make it easy to combine and change databases. A problem related to Reference Manager's single-database limitation is that people tend to create very large databases, which causes the program to bog down.
For a booklet, published in 1993, comparing features of three Macintosh bibliographic-management programs, Reference Manager, EndNote Plus, and Bookends Pro, contact DCRT's Technical Information Office (fax: 402-0637; e-mail: firstname.lastname@example.org).