Life is not easy these days. The costs of adding an electronic versions of an STM serial to which a library already subscribes can vary from nothing to an additional 30 percent (and I am sure there are examples of even higher prices). Seemingly the subscribing library can do very little about the price for the selected list of journals the library actually wants to receive from the publisher. What is worse, most publishers have yet to make adequate provision for ensuring the archival quality of their electronic journals, much less offering to make them available for more than two to five years. Adding insult to injury is the poor quality of the majority of the offerings which are offered as electronic journals. Many electronic journals do not have adequate links to references, significant electronic features, or the structure, layout and navigational aids which make them easy to read on the computer screen.

Admittedly, we are in the midst of a major revolution, and it takes time to change massive publishing operations. Still, the progress towards the efficient production of good electronic journals seems to be taking a long time, and it is the library budgets which are paying the price. Consortia pricing offers and access to bundles of electronic journals to which a library does not really care to subscribe do not solve the underlying problem which is the high prices being charged by some publishers for electronic journals.

It is not clear that cost have to rise. Of course, producing an electronic journal using procedures developed for the paper-centric past is going to be expensive, and the resulting journal is generally not well designed for easy archival maintenance. However, changing the publishing process into an electronic-centric one can offer a modicum of price reduction, pave the way for effective maintenance of the electronic archives and make it easy to add reference links and other features which make electronic journals easier to use and more effective in transmitting information from the author to the reader.

Publishers who say that electronic journals cost more because they have the added cost of producing the electronic files are probably still producing the standard paper journal first and then adding on the electronic features. Under this scenario, yes, producing the electronic version will add costs, and the electronic products will probably be less than optimal. Our experience at the American Astronomical Society show that a different scenario can be achieved.

Five years ago, the American Astronomical Society forged a partnership with the University of Chicago Press to bring our journals electronic in the most effective way. Our goal was to change our publishing process so that we produce first an electronic, archival quality master copy of our journals, and then derive the publicly displayed screen versions as well as the paper version relatively inexpensively from that master copy. The master copy is heavily tagged in SGML which is the secret to making this possible. We believed that by standing the old production process on its head and tailoring the new system to the electronic world, we could produce an electronic version of our journals which would come out more quickly, be effective vehicles for transferring information (i.e. be of high quality), and cost no more than our paper journal originally cost.

It has not been an easy task to revamp our entire publishing process for our 30,000 journal pages per year while still continuing to put out an issue every ten days. One member of our team, Evan Owens of the University of Chicago Press likens it to changing the seats of a full 747 while it is in flight. Now, five years after we started, we are able to say with certainty that we have succeeded. In the process we are reaping benefits in addition to cost savings, which are more important than we realized at the time.

• We have proven that the effective use of electronic tools and automation can result in reducing the costs associated with producing the prime electronic copy.



• The paper version can, indeed, be derived almost automatically from our electronic master copy. The only cost associated with producing paper is the marginal cost of printing, binding and mailing.



• The electronic versions, of which there must be two, a version for reading on the screen (e.g. HTML with abundant links and electronic features) and a version for local printout (e.g. PDF), which can be derived relatively easily from the electronic master copy.



• The electronic master copy, if done well, is able to serve as the robust electronic archival copy. Such a well-tagged copy can be updated periodically, at very little cost, to take advantage of advances in both technology and standards. The content remains unchanged, but the public electronic version can be updated to remain compatible with the advances in browsers and other access technology

We have found that by using the electronic-first philosophy we could offer the electronic-only subscription as our prime product at 70-80 percent of the present cost of the paper subscription. A paper subscription, should the library desire it, and most will for a while -- maybe for ever -- can be added for an additional 20-25 percent. The total cost of the paper plus electronic combination should not have to exceed the present cost of the paper subscription produced using the old process. In our own case the preliminary indications are that in inflation corrected dollars, our total production costs have probably decreased slightly over the last five years. We will have better figures in a few months after our fiscal year closes.

But price is not everything. The archival management advantages inherent in our approach are in some sense even more important.

• First, we have to realize that an electronic archive needs constant tending and updating to keep up with evolving standards and new technology. We are very confident in our ability to preserve the master copy of our electronic journal indefinitely. The robust format of the electronic master copy allows it to be updated and managed and translated into any new standard so as to ensure the permanence of the electronic archive. We now believe that this will be able to be done within the operating budget of the journal. The key is that all updates can be done by automatic computer scripts and programs which do not require human intervention beyond the development and checking of the scripts themselves, which is not a large job.



• Second, we have a way to maintain the technological currency of our journal. As browser technology advances (or at least changes) all issues of the screen version of the entire journal, from the time it first went electronic, can be updated by automatic scripts and programs so as to incorporate the new capabilities. At the very least, such updates will allow the material to remain accessible, even if all the new browser capabilities can not be incorporated. This keeps the journal (including the links, the videos, the live math, the computer readable tables and all the other electronic-only features) alive and useful as technology changes.



• Third, the philosophy of maintaining an underlying electronic database and deriving the screen version from it can be transformed into a process of deriving the screen version on the fly. There are two reasons to do this; to make pages customized to the user, and to make pages customized to a specific browser. If the reader happens to be in Japan, the pages can be a Japanese translation, a feature already offered by Nature. Also the pages can be customized for the browser being used - to take advantage of the full suite of capabilities. In the ideal world, the browsers would not differ in major ways. Alas, in their attempt to capture and retain market share, Netscape and Microsoft are adding more and more incompatible features. To take advantage of these special features, electronic publishers have to customize their pages to the individual browsers. Even in the world of academic publishing, the days where we can simply serve up static pages is coming to a close.

In summary, our experience has given us an insight into the advantages of re-engineering the publishing process. First, we have found it possible to produce both the paper and electronic versions of our journal at a total cost equal to or slightly below our total costs five years ago. But, holding the line on price is less important than achieving the robust archival capability and the flexibility to incorporate new features and tailor the form of the output which now become possible.

Along the lines of my previous article, ("It is Time to Become Discriminating Consumers," P. B. Boyce, ATG, v. 5 # 9, p. 86, Nov 1997) it seems to me that, as users and "keepers of knowledge" the library community should be insisting on the same standards of quality in the electronic journals as you do for paper products. Would you actually pay for a paper journal whose printing was so bad that it was struggle to read and whose paper was sure to crumble within five years? Why should you settle for less than the highest quality in presentation and features? And how can you think of accepting anything less than archival quality formats in the electronic products you license?

Under ideal conditions all of this should be possible at the same price you paid for paper five years ago. But I have been discussing the costs, not the subscription price. There is a difference, of course. Publishers set the actual subscription prices based upon a number of factors besides the actual costs. In our case, for instance, where half our journal income is based upon collecting page charges from the author, we have chosen to raise subscription prices in order to reduce page charges somewhat. But the underlying point remains. Subscribers to electronic journals should investigate the archival quality of the electronic material they are licensing, and they need not accept at face value that adding an electronic edition must necessarily cost more.