Vancouver, British Columbia
May 15-17, 1996
James G. Neal, Eisenhower Library Director
Johns Hopkins University
I’ve been asked to talk about Project MUSE, which I have come to refer to as an adventure in scholarly publishing. I think it’s important to remember what Mark Twain said when telegraph service was established between the coasts of the United States. He said, “Maine has contacted San Francisco, but Maine has nothing to say to San Francisco.” That is very much at the root of what we are trying to do with the Project MUSE initiative: to give that developing international network meaning and relevance.
We also recognize that not all the people on our campus understand and can properly work with the MUSE solution, particularly our scientists and engineers. Our acronym, Many Users in Scholarly Exchange, indicates that the database is meant to allow you to use what you will.
My objectives here this morning are to describe the origins and objectives of the project; to summarize our current status and some of our short-term goals for expanding and improving the project; and to share some of our preliminary findings and observations.
This project was launched several years ago with a set of objectives. The first clear objective was to make the 42 journals of the Johns Hopkins Press available electronically and to create a successful model for electronic scholarly publishing.
One of the primary interests we had was to increase access while decreasing serials costs at the same time. We wanted to provide not only access to what was in the journals but to add some significant value to searching those information sources in the academic community.
We particularly wanted to look at network-based, non-proprietary interfaces to both take advantage of what was being developed and to avoid investing too heavily in specialized proprietary products. We wanted to maximize the ability of a campus to distribute and duplicate scholarly information, so from the outset we’ve been very focused on intellectual property issues, which I’ll later speak to in a little more detail. And we wanted to actively involve librarians in designing electronic scholarly communication, to develop the partnership between the press and the library. That has been a centerpiece of the Project MUSE initiative right from the start.
As we have moved from prototype to production, we have tried to build upon this cooperation between the library, the press, and the academic computing organization at Hopkins. It’s important to recognize the early work that was done by Scott Bennett, the previous library director at Hopkins, and Jack Goellner, who was the director of the press during the conception and the successful launching of this project. Having that proper framework set up early is very much at the root of our success.
We were also successful because of the availability of three-year’s funding from the Mellon Foundation and from the National Endowment for the Humanities to enable us to put this infrastructure in place and move it forward.
We wanted, at the outset, to build a library-based subscription model that would build on the traditional ways libraries gather journal-based information. We wanted to design HTML documents so that as many browsers as possible could use the documents easily. We’ve recognized that we’re dealing with a volatile environment and we wanted to support all the systems that institutions and users would have.
We wanted to maximize the ability to search both individual serial titles as well as the entire database, the whole corpus of the 40-plus titles. Searching the entire database includes access to the table of contents or the titles with key-word and Boolean capabilities. We wanted to maximize the ability to get into this database and secure excellent results. We also wanted to be able to support, in any university environment, an unlimited number of simultaneous users. It’s these ideas, as we have moved from prototype to production, that affected some of the technical and service decisions that we have made.
Throughout this process we have been very much influenced by some of the ideas that came out of the AAU/ARL initiative. As some of you may remember, there was an important effort in the early work of the AAU/ARL initiative to write out a set of functional and performance attributes for this technology. It’s important to recognize that this collaboration in our ARL community is influencing the thinking that’s going on in some of these electronic scholarly publishing projects. We want to be sure that we have ease of use and that we are providing information that is authentic. We want to have a process that enables, through eligibility devices, people on campus to use it. We want to control costs. We want to be innovative. These principles have always been at the forefront as we have planned the future development of this project.
I mentioned the importance that we have placed on enabling expanded search and use capabilities. As I said, this includes the ability to search an individual title, the ability to search subsets for groups of titles, or the ability to search the entire database. Included is the ability to search tables of contents, the full text of all the articles, or assigned subject headings. We have contracted out the subject heading development for each of the articles in the journals, and this is coordinated by the library’s cataloguing department. Each of these search strategies have both keyword and Boolean capabilities.
The search results that come back to the users— faculty and students— the journal title, the volume, the issue, the author, and article title for all the items that meet the search terms. These results are then provided in a relevance-ranked order, so there is some commonality between how MUSE and how other web-based sources that people are using respond. Further, we have built hypertext links into the table of contents, the endnotes, the author biographies, and the illustrations that are part of the articles.
We have tried to design the text for easy on-screen reading, and, as much as possible, to enable people to consume this information by developing a comfort level with the screen rather than automatically printing. For instance, we are providing illustrations larger than those in the print version and, increasingly, those illustrations are being communicated in color.
We have, however, also tried to provide the ability to download and print the information easily. We are finding that we can deliver the electronic version of the articles as many as four weeks earlier than the print version actually enters the library. Speed of access, combined with the searchability of the information, creates a very powerful package of user support.
Another thing that we have been very focused on in developing and implementing this project are issues of enabling fair use. The library involvement in this project is most evidenced in the service side and in the fair use side. We are applying no limits on the number of simultaneous authorized campus users may be allowed. As many people as can get in can use it. People can download, save, and print articles for personal use. They can be applied to multiple-copy distribution for classes, as long as they’re not sold for this purpose. People can print out articles and use that hard copy for inclusion in material selections. An unlimited number of copies can be put on reserve, whether they are paper or electronic. Notices can be placed on campus, in whatever network focus of interest, to inform users that these sources are available.
We are enabling subscribing libraries to download and save the materials locally to main servers, archiving it on CD-ROM or tape backup. The principle that we have tried to apply here— have argued quite strenuously for— that libraries own the materials to which they have subscribed. So if a subscription ends, ownership of the materials during the subscribing years is retained. We are also enabling interlibrary loan of articles based on the CONTU guidelines using facsimile page systems, electronic delivery, faxing, or photocopying. The only restriction that is currently in place is the electronic transmission of articles outside of the subscribing campus. We know that distinction is beginning to blur. This has been a very difficult issue, and one that the library and press sides of the project have strenuously debated. We are monitoring the situation very carefully and we’re trying to weigh the interests of the journal sponsors, who have fundamental concerns in this area, against the needs and interests of the library subscribers. But I think you can see that there has been enormous progress made in developing and implementing this project and supporting the fair use and other rights of the subscribing libraries.
We’re also trying to lay out a flexible subscription model with a whole variety of options. One thing we wanted to do right from the outset is not link print and electronic subscriptions. You do not have to have a print subscription in order to get the electronic version. As many of you know, we have a free trial access— open searching capability, for private use— the databases. Certainly from the library’s point of view, it has been an opportunity for worldwide exposure to Project MUSE. And from the publishers’ point of view, it has been an opportunity to reach a very important and expanding market of interest.
In terms of the economics, we had no basis on which to make projections. We did some preliminary analysis of the costs we thought we would experience and we started out by pricing the electronic versions of the journals 10 percent less than the print price. This enabled full domain institutional access. We also decided that, for the institution that subscribes to both the print and electronic versions, the price for both would be 130 percent of the print price. In addition, we have laid out very rigorous individual library as well as multi-campus consortial pricing plans. We are seeing enormous interest in consortial subscriptions, both from the current customer base for the press titles as well as from many new customers who are coming forward.
There is a transition pricing plan for 1996. We are trying to reduce the bureaucracy of getting campuses online by providing access through domain name and IP addresses. We have now over 250 subscribers to the current versions of 1996 online journals, and over 1,000 additional individual libraries are in discussion with us regarding a basis for access to use. So it’s a very successful enterprise from both the presses’ and libraries’ points of view.
We see two reasons for libraries to come on board at this point. First, libraries are indicating to us that they want to demonstrate a vote of confidence for a project like MUSE because it is university based and there is library involvement. But perhaps more importantly, there is an interest in providing a positive campus experience with electronic publications, and libraries recognize MUSE as a package of journals that have a high reputation and for which good service and intellectual property provisions have been made.
I just wanted to talk a little bit about some technical considerations. We have determined, obviously, that the MUSE titles are best viewed with a graphical browser like Netscape or Mosaic, but we continue to support text-only browsers by Lynx. Our experience is that currently about 90 percent of the users are coming in with some type of Netscape browser, about five percent with Mosaic, about two percent with Lynx, and about three percent are coming in with as many as 70 other browsers. You can see the complexity of the environment in which we work.
We are currently seeking to conform to HTML 3.0 type specifications and we are indexing each journal with SWISH, the Simple Web Indexing System for Humans. We have made a major conversion change. We were looking and working initially with postscript HTML processes. Now we are trying to work with the typesetters themselves at the journals to try to ease the transition from the print base to the electronic distribution. That has produced some enormous economies in the conversion process. We then apply the in-house MUSE scripts, which upgrade the overall quality of the electronic text and content for each of the HTML files.
We have put together some very important per-article and per-issue production checklists and tracking capabilities, which is really facilitating the conversion, preparation, and proofing of the HTML documents that is done by the large student staff that has been integrated into the project. This typesetting to production efficiency and these types of management capabilities are demonstrating, at least for us, that the electronic cost of production is less, in fact, than the print cost of production.
We have also found that the initial investments we needed to make in order to get this up and running and to move from experimentation to production were less than we anticipated. Also, we did not expect to be at this point in terms of the number of journals we would have online and the number of subscribers that we would be supporting. Actually, we didn’t expect to have what we’re experiencing today to be in place until 1997.
We’re investigating the Text Encoding Initiative and, of course as we’ve all indicated, the SGML content tagging capabilities. We provide start-up technical support on the server side, but because of the diversity of the university campus environments and the number of browsers that individuals are coming in with, we do not provide that support on the browser side. We are, however, tracking the network platforms, the browsers and their versions, of the visitors who are coming in, as well as of the users. As I indicated, it’s predominantly Netscape and Mosaic. We have implemented a very dynamic allocation of the most appropriate HTML bandwidth pages for specific browsers or browser classes. So that when someone comes in, the system automatically responds to that browser and presents the information in the most effective and viewable way for that particular browser.
We are collecting an enormous amount of statistics, and subscribing libraries have automatic and immediate access to these statistics for their users. Article hits, image hits, hits on the table of contents, the number of times each journal is used, the ranking of the articles used, and the ranking of subscriber activity are all measured. So there’s an extraordinary database of information that is being developed here, with implications on the library, consortial, and national levels.
We have created a robust cyber-IP database with robotic investigations and nightly updates. So making sure that the proper people and institutions are authorized is being increasingly managed by machine instead of through staff intervention. Because we’re running this on a nightly basis, it’s really built in a high level of efficiency and support for subscribing libraries. We are also getting regular reports on access denials and other responses that might indicate that an error has occurred. Those are automatically logged and sent to the appropriate staff member in the Project MUSE process for investigation.
We have also purchased an open-market secure commerce server that we think will be important down the line as we implement individual subscriptions and electronic document delivery. Our overall technical goals have been to use, as much as possible, standard non-proprietary software, to build in tools that enhance user support and that advance innovation. As the other speakers have indicated, this is a very complex and regularly changing technical environment, and our ability to be agile in applying some of these new technical capabilities will be critical to our success.
What are we trying to do in the short term? We’re trying to complete the implementation of what is now 40 Hopkins Press titles. We have an ongoing dialogue with the journal sponsors about the order in which these will move forward and we are anticipating that by the end of 1996 we will have all 40 journals in production, with all 1995 and 1996 issues in place.
We are seeing a rapid growth in the subscription base and we think that will continue through 1996. We had a preliminary goal of 500 subscriptions by the end of this year, and we think that there will probably be in excess of 1,000. We are now exploring options for expanding access to back files prior to 1995. We’re looking at onshore and offshore, local and contract-out capabilities. We think that expanding the database beyond the current two years is very important.
Although this is a dynamic subscription environment, we’re beginning to get a better handle on the economics of the scholarly publishing model. Our early projections of pricing electronic versions of journals at 90% of the print subscription price will probably be on target for these initial years. But only when the subscriber base expands and we apply technical efficiencies will we really get a firmer handle on what the economics of this will look like. The major concern, of course, that many of the journal sponsors bring is the ability to create a tenable income stream. So, over the next year we will look at how that factors into the overall economics of this process. There needs to be an acceptance of this medium for distribution and use in order for a project like MUSE to move forward, since we don’t own, if you will, all the journals that we distribute.
We are in the process of implementing a very substantial online questionnaire which we are prompting all of the users of MUSE to take advantage of and we hope to gather a lot of user feedback over the course of the next year as more and more people provide us with that information. These very extensive use data are being collected and analyzed and is, as I indicated, provided to subscribing libraries. We are very interested in seeing how this project will influence collection development, management, and management behaviors.
We are promoting author awareness and creativity, convening not only user focus groups but also convening focus groups of authors so that they can begin to look at the audio and video capabilities and understand what it means when their products are in this type of environment. One way we are monitoring progress in this area is by paying attention to developments affecting the SGML content tagging, which we believe has some real potential in terms of enhancing certain types of titles and information. Also, we will look at implementing hypertext links to the references and datasets so that when one reads an article in Project MUSE and comes to the references at the end, there will be the ability to automatically move to the text of those referenced articles. This involves heavy negotiation with other journal publishers, one of the really exciting aspects of the project. The other, of course, is that in many journals there will be a set of data that is direcly linked to the findings. We want to be able to link users into data sets that are located at other sites, as well.
We now have to think in terms of the long run on how this project will continue as we move into very high production and an expanded environment. There are print journal publishers approaching us and asking about the prospective distribution of their titles in electronic form through MUSE, and we have an increasing number of people who are putting up or who are conceiving electronic journals and who want to use MUSE as the platform. That will enable us to create subject clusters so that individuals can subscribe solely to higher-education titles, mathematics titles, or literary-analysis titles, for example. We are also looking at some editorial limitation, what we’re calling Product Alpha, where we try to get rid of some of the current conventions that are in print journal literature. And, as I mentioned, we’re looking at expanding to include individual subscriptions and document delivery of articles.
My last point will be some preliminary observations. Subscriptions are predominantly to the entire MUSE database; we have very few libraries subscribing to individual titles. We are also finding that we are working with few individual libraries. In most cases, we are working with consortiums that are coming forward to negotiate access for all of their libraries, and, in one case, there is negotiation of a national license agreement with another country.
Libraries are adding hot links to MUSE from their online catalogues when that is possible. There is expanding interest in contacts with current print journals as well as with electronic journal producers about MUSE as an outlet for their titles. This has produced a very significant new customer base for Hopkins Press. Many of the individuals, particularly those coming in through consortiums, have no, or very few, current subscriptions to the Press’ titles. In two of the large consortiums that we are currently implementing the MUSE Project, the average number of titles currently subscribed to in print in one consortium was eight, and 13 in the other. Remember, there are 40. So what we can present to the Press is a massive increase in their customer base, which still needs to be evaluated in terms of the long-term economics of this scholarly publishing model. We are also seeing extraordinary expansion in international interest. We have had hits on the Project MUSE home page, http://muse.jhu.edu/muse.html, from over 70 countries. So we’re getting a lot of exposure to press titles that probably would not have been possible in the print environment.
There is concern coming from the journal sponsors about the future identity of the journal title, particularly when people are subscribing to MUSE. What happens to the journal title in that context? First, there is a need on the press side for technical support and customer service personnel that was not needed in the same way in the past. This will eventually have to be factored into the overall economic analysis about new types of and perhaps more staff involved in supporting this kind of access. Second, we also have recognized the need among the journal users for familiar elements. So, if you have looked at MUSE, you know that we have retained page numbers, for example. That’s a small example of how we are building bridges of familiarity between the print and the electronic environments.
Let me conclude by drawing upon, again, the important work of AAU and ARL. This has been a very important for us as we have proceeded with Project MUSE. We are trying to foster a competitive market for scholarly publishing by providing realistic alternatives to the prevailing commercial publishing options. The IScAN proposal that we have been asked to look at by the ARL Board and the types of initiatives that have been presented here this morning indicate that there is considerable energy moving forward toward this goal. We need to develop policies of intellectual property management that emphasize broad and easy distribution and reuse of material. What I’ve tried to demonstrate is that Project MUSE is actively trying to define in new ways what intellectual property management and fair use mean in this electronic journal and publishing environment.
Innovative applications of available technologies that enrich and expand the available means for distribution of this research and scholarship are necessary. Through the types of searchability and accessibility that projects like MUSE present, I believe that we are supporting AAU and ARL interests. We want to assure that the channels of scholarly communication sustain quality material. I believe that as we bring new electronic journals into this Hopkins Press environment there will be the leveraging of that traditional quality that university presses bring.
Lastly, we want to make sure that there is permanent archiving of these research publications in digital formats. That, as the Commission/RLG report has indicated, represents a challenge for us in the next couple of years.
Thank you.
Copyright © 1998 by James G. Neal