During the life of ANON there have been guest contributions from a number of sectors -- educaton, professional societies, geological surveys. One missing sector is that of the publishing companies. This month I have asked Bas van der Koek, Publisher, Earth Sciences, Elsevier to share some of his ideas and concerns with the readers.

Introduction In 1663 the Journal de Sçavants, the first scholarly journal, was launched, a few months later followed by the still existing Philosophical Transactions of the Royal Society of London. Since then, very little has changed in the scholarly publishing industry. However, in the last few years, changes abound, fuelled by the advent of World Wide Web. Publishers realise that on the one hand the market is demanding electronic publishing, and that on the other hand they can do a better job with more efficient production and distribution. How have publishers, such as Elsevier Science, prepared themselves for this? What is the long-term outlook?

Introduction

What are publishers doing now?

Often headed by the large learned societies, publishers have had many initiatives on the Web. These include:

1. Alerting services on the Web: these are mostly free of charge, and include lists of contents and often abstracts. Surrounding infrastructure such as a search engines are also provided. Good examples can be found on the AGU web pages. The National Centre for Petroleum Geology and Geophysics offers and excellent overview of the web-presence of earth science journals. Publishers hope that by providing abstracts readers be prompted to use their journals more often, thereby increasing readership and exposure.

2. Publish "unprintable" items on the web, such as movies, sound, or computer programmes which are connected with articles which are conventionally published. A good example is the programme library of this journal , which is maintained by the editors themselves. Our firm publishes sound files with the journal Speech Communication. The Geological Society of America maintain the GSA Data Repository which contains mostly additional, supplementary data, such as tables with detailed measurements.

4. "Electrified" existing journals. Springer ( has most of their journals in pdf format which basically offers readers the possibility of reading (or printing) the articles in exactly the same form as they are printed in their respective journals.

5. Real electronic journals in HTML format . Example: the existing journal Earth Planetary Science Letters (this has features such as electronic datasets, and also provides the abstracts of the literature references). New journals include Earth Interactions.

Is all this really any better?

Above services do have their nice aspects, but on the whole I do not really believe that they will seriously undermine the popularity of a printed journal issue. Yes, it is convenient not to go to the library anymore to read a journal, but the amount of journals full-text available from the desktop is still very low. And as most WWW publications imitate the printed product, how can one improve on the magnificent interface offered by the print medium? This has developed from the ancient times of clay tablets and book scrolls and now the interface of a paper issue has page numbers, a list of contents, keywords indexes, but also ease of reading, convenience of flicking through pages, and portability. There is simply no comparison with a computer screen, which is not portable, awkwardly hard to read, even harder to browse, may have inexorably slow connection times, and from time to time a breakdown. Other major concerns are the stability of electronic publications (paper has proven to be quite ärchivable", but how about electronic formats? Is a pdf file still legible with the equipment that is to be used in 2097?)

Outlook

In my view, a successful electronic product should be complete, easy-to-use, and make maximal use of features offered by web-technology. I believe that the following points are essential ingredients for this:

1. Complete offering and easy usage: scientists are more likely to go to electronic literature, once this offers a fairly complete overview. Thus, publishers should link their services using the same standards for representation and interface.

2. To achieve eternal preservation, articles should be stored in a generic format. Many publishers (including IEEE, ACM and Elsevier) adhere to SGML (Standard General Mark-up Language), which allows them to represent articles in a contemporary format (at this time this is conventional print, pdf and HTML3.0; five years ago it was postscript and HTML1.0). With SGML, the text is exhaustively coded (tagged). For instance, literature references are tagged, detailing author name, journal, publication date and so on. This allows a publisher to link references to abstracts (from GeoBase or GeoRef) automatically, and even to link to the referenced article.

3. Good searching methods -- not a search engine based on inanimate-like boolean logic, but a "fuzzy logic" search mechanism, using natural language as well as a good thesaurus. Who knows? Maybe finding the information becomes more important than the information itself.

4. Link the primary literature with secondary literature Wouldn't it be nice to do searches on GeoBase or GeoRef, and be hyperlinked immediately with the full-text of articles resulting from your search? And conversely, while reading and article, being able to click on the literature references and go to the abstract (or better still, to the article in full text -- if within the database)?

5. Use the medium to the fullest.
*Multimedia files, such as animations, movies and sound; Java code
*"Living" articles: articles that can be revised and updated after publication continuously
*computer programme code, plus test runs.

A journal publication which features a number of the above points is Earth & Planetary Science Letters Online, which boasts a link with a secondary database, has HTML and pdf articles generated from SGML code, as well as multimedia files (called datasets).

Putting one journal on the web is easy, but for all of Elsevierıs 1,100 journal titles is a different story! A uniform tagging system needs to be agreed upon, production of impeccable SGML code needs to be arranged with the many subcontracting typesetters, and special hardware configuration (called "electronic warehouse") has been constructed to store the 1,000,000 articles published annually by Elsevier. Our firm have now 100 life sciences journals available in HTML/pdf in the project Science Direct, which have a direct link with EMBASE, Elsevierıs abstracting/indexing system in the life sciences. Next year, we hope to expand this to all our 350 life science journals, and I hope Earth Science (with links to the abstracting/indexing service Geobase) will soon be included.

The traditional added values of the publishing industry, such as composing, printing and distributing of articles are vanishing. And although paper publication is hard to beat, I am sure that the publishing industry is developing an electronic publication system which really will makes an impact on how scientists will consume the earth science literature in the future.

Any reactions to the above will be highly welcome.

Since September 7, 1998

Return to ANON Columns