As explained in Eos last summer, starting 1 January 2002, AGU journals dispensed with sequential page numbers. (See the item by Sam Bowring at www.agu.org/pubs/e_publishing.) Some members have questioned this approach and so we, the Publications Committee, revisited the question via an e-mail discussion this spring. After carefully thinking about the issues once more, we remain convinced that the course charted by AGU is correct.

AGU is not alone in dropping sequential page numbering. For example, the American Physical Society dropped page numbers a couple of years ago when they declared their electronic version to be the journal of record. APS uses a 6-digit article number and instructs authors to cite that number in lieu of page numbers.

AGU is using the Digital Object Identifier (DOI) in place of page numbers in our citation style. It has been suggested that this usage is "highly eccentric" and "out of step with general trends in online publishing." Our alternative view is that AGU has taken a bold step ahead of almost everyone else.

We believe, however, that in the near future, more journals will move to the use of the DOI and that it will become clear that AGU's practice is not out of step, but rather, defines the trend for the industry. Already, Nature has adopted the DOI as part of its citation. The following is a quote from the Nature Web site: "Given the utility of the DOI in locating an online publication in the future, we encourage you to use DOIs in your reference citations." Science also instructs authors to include the DOI when citing articles from that publication.

For AGU, the electronic version is the journal of record, and the Union is assuring the long-term archiving of this version. The html is the online representation of the journal of record. The DOI will be essential for locating items in the record in the future. AGU moved from print to electronic so that non-printable material would be an integral part of the record of science, and not simply supplementary to it. Even though other publishers are including non-printable material on their Web sites, it is not a part of the formal record as long as their print version is the journal of record.

We believe that the method of citation chosen by AGU is rather clear and that users of the print version will not have difficulty in locating an article after a little practice. The AGU-recommended citation style (see accompanying example in text box) includes the volume and the issue number, and as discussed in the 13 August issue of Eos a citation number. The issue number is an aid for those using the print journal. Once the user gets to the printed journal, he/she finds an author index and a DOI index for that issue. (The author index will be cumulative starting with the August issues.) These indexes indicate the article order within that printed issue. AGU journals assign an issue number and the citation number at the time of publication. For special sections, the issue number may be several months in the future. The issue number and the citation number are keys to finding an article on a shelf of printed journals. Both numbers appear on the spine of the printed version of the journals; the issue number has always been on the spine, and the citation number will join the issue number beginning in August.

It has been suggested that we could search for interim "solutions" that would allow sequential page numbers to co-exist with the DOI. We have considered this issue and find that the costs for AGU of doing so far outweigh the benefits. Here are two ideas and our reasons for discarding them.

1. We could have sequential page numbers in the printed journal if we did not provide a pdf online until the journal is printed. (Otherwise, we have to pay the cost for remaking all of the pages just to get the page numbers into them for the printing process, an additional cost that we believe is not justified.) Delaying the pdf strikes us as a terrible idea.

2. We could go the route of Nature and cite the paper one way before printing and another way after. The Publications Committee sees this as an option that potentially causes confusion rather than enlightenment. We chose the DOI as a consistent and persistent way of citing an article from the moment of publication. The DOI is an international standard designed exactly for the purpose of citing electronic material in a way that persists regardless of changes in technology or changes in the server on which a particular file resides.

In summary, it is our opinion that we cannot simply stagger along attempting to keep up with changes by applying short-term, inadequate fixes. We believe that the move to the DOI as a citation standard may put AGU ahead of the curve, but the rest of the world will soon catch up and we all will become used to the new way of citing articles. We hope that all members will stay the course with AGU, even though some may remain skeptical. At the pace of advancements in electronic publishing, the results of the "experiment" will be known to all of us in very short order.

Author

George M. Hornberger, Chair
Publications Committee

© 2002 AGU