PROPOSAL FOR A DIGITAL PRESERVATION COLLABORATION

1. ENDURING ACCESS TO ELECTRONIC JOURNALS. Authors, publishers, and librarians all feel it is imperative that assured long-term access to electronic journals be provided. The inherent mutability of information technology, a rapidly changing marketplace for electronic publishing, and the absence of any well-established preservation practice for electronic information make the prospect for long-term access most doubtful. Until we have successfully addressed this issue, scholarly communication will remain bound to print and paper because with these formats we can assure long-term access to traditionally published scholarly information. This is a costly and dysfunctional bondage, where electronic publication otherwise serves scholarly communication decidedly better.

The Yale Library and Elsevier Science regard the long-term access to electronic journals as mission critical. We feel that both organizations have vitally important, complementary capabilities to bring to a digital archives project. Our experience in negotiating licenses leaves us feeling confident that we respect one another's institutional imperatives and value the human talent in both organizations that must be engaged in an archival project. We believe we will be able to manage genuine differences in our organizational missions and cultures in a way that will allow us to work in close collaboration on our shared concern for digital archiving.

Assured access to electronic journals embraces a variety of activities that might usefully be grouped in three categories, each distinctively driven by:

• The circumstances of ordinary commerce. Included here are access arrangements (such as mirror sites) designed to deal with ordinary telecommunications, time zone, and redundancy needs. This category also includes arrangements for responding to extraordinary emergencies (such as massive equipment failure, damage done by hackers, etc.).
• Likely changes in publishing. The Yale Library and Elsevier Science recognize that some arrangement must be made to ensure enduring access to electronic content in the face of changing commercial circumstances. What should be done when some substantial part of a publisher's content is no longer commercially viable (the digital equivalent of going "out of print")? when a publisher sells a particular journal, ceases publication of one or more titles, bundles material in significantly different ways, or disengages from a certain line of business? when a publisher is sold or goes out of business? Such changes in a company's commercial circumstances pose grave threats to enduring access to its publications. In an enterprise as volatile as publishing now is, we should expect many such changes in any 10-20 year period.
• Traditional archival needs. In the very long term (100 years+), most readers will address questions to an archive of Elsevier Science publishers unlike the questions of today's practicing scientist. These questions will normally be of a historical character, and they will be addressed to an archive structured in ways that are especially responsive to historical inquiries. These questions will reflect the intellectual paradigms of the future, which can hardly be predicted today.

The term "archiving" has been applied somewhat indiscriminately to all of the situations where assured long-term access to electronic journals is needed. Elsevier Science provides for timely, uninterrupted access to its electronic publications as a part of its ordinary business. The publisher has its own telecommunications capabilities and back-up arrangements in place and is exploring additional arrangements with national libraries and local-load customers to guard against more damaging emergencies. Recognizing these arrangements, the Yale Library and Elsevier Science propose to take action on the two other access needs identified above.

We seek funds from the Andrew W. Mellon Foundation for a year-long effort to plan our archive in detail and to create the test-bed computing capabilities needed for that planning.

2. DEFINING THE SCOPE OF A COLLABORATIVE PROJECT. The Yale Library and Elsevier Science intend to work collaboratively in the following ways:

• The library and Elsevier Science both wish to understand the ordinary commercial life cycle of scientific journal articles so that we might create an archive that protects readers from the failures of access that are likely to occur in that cycle. We believe that the competitive, commercial marketplace can be a highly efficient way of providing access both to individual articles and aggregations of them (say in journals or in still larger databases). We also believe there is a time when the marketplace will no longer support access to some part of a publisher's output. This point in time will vary immensely from article to article, from journal title to journal title, from discipline to discipline, etc., and neither publishers nor librarians yet know enough about the online environment to predict confidently when this time will come. But everything in our experience suggests that continuing always to burden commercially viable content with content that is not, or that is only marginally viable, robs us of the economic efficiencies of the marketplace. So burdened, a publisher is likely to change its business plan or, in the extreme case, go out of business. These are points of great risk for scholarly communication, as the ordinary uses of the threatened journals may become unsupported. We must develop an archival practice that guards against this danger.
• The library is interested in working with Elsevier Science to support ordinary archival uses of that company's journals published electronically. Responding to well-established patterns in the use of archival collections, our archive will be designed to support two principal kinds of inquiry, both of them significantly unlike the needs of the practicing scientists that Elsevier Science journals are created to support. These are (1) inquiries into the scholarly, business, technical, and policy environments within which Elsevier conducted its publishing business; and (2) inquiries that treat archival information in transformative ways. Inquiries of the former sort usually draw on a wide variety of resources and are sometimes possible only after confidential or commercially sensitive information in an archive can be opened for study. Inquiries of the latter sort usually address subject or disciplinary interests unlike those that originally prompted the publication of the material in the archive. Such inquiries typically require some reconfiguration, re-assembly, or re-purposing of the archival information (often in conjunction with other archival holdings). There is no way confidently to predict what these archival inquiries will be, who will make them, and when they will be made. The most likely prediction about these archival uses, judged by other archives, is that they will be relatively infrequent initially but will increase over time as the scholarly value and integrity of the archive increases.

3. DISTINGUISHING CONCEPUTALLY BETWEEN CONTENT AND FUNCTIONALITY. In the print world, intellectual content and the physical presentation of that content (including, for instance, indexes and elements of graphic design) are immutably linked in discrete, time-bound publications. Whole intellectual disciplines, such as descriptive bibliography, have developed around this fixed relationship between content and its physical embodiment (the elements of which are sometimes referred to as paratext features). In the digital world, no such immutable links exist. Indeed, readers positively expect the presentation of content to change, in the form of new system functionalities that enhance use. But at the same time, readers are most uneasy with the thought that the intellectual content of electronic publications might change in ways unknown to them. Readers insist on the authority and authenticity of digital content but welcome changes in functionality.

The Yale Library and Elsevier Science, observing this striking difference in reader behaviors in the print and digital worlds, wish to test the viability of archival services based on a fundamental distinction between the "content" of the scientific articles Elsevier publishes and the "functionality" provided to readers by Elsevier at any given point in time. We believe this distinction may be critically important in setting and meeting reader expectations about digital archives, in devising technical strategies for meeting those expectations, and in controlling the long-term costs of archival services. We have therefore made the distinction a central feature of the exploratory planning we are proposing to the Mellon Foundation.

We define content as the data and discourse about the data that authors submit to the publisher, along with other discourse-related values added by the publisher's editorial process (e.g., revisions prompted by peer review, copy-editing). We define functionality as a set of further value-adding activities that do not have a major impact on the reader's ability to understand content. So understood, functionality embraces a very wide spectrum of things, from graphic design to search engines, and from branding or aggregating practices to the provision of links among related materials. Functionality has most to do with reader productivity (e.g., the ease and efficiency, or even the pleasure of use). Functionality is not linked to individual content but is pervasive to large bodies of content. In posing this distinction, we recognize that the boundary between content and functionality may not be crystal clear in every circumstance. It is likely to vary at any given time as one moves from one disciplinary literature to another, from one publisher to another, and from one journal to another. It is also likely to change over time, as technology evolves. While we recognize this significant variability, we also believe that the distinction is inherent to digital technology and is likely to figure centrally in any viable archival strategy.

More specifically, we believe an archive of electronic journals must unquestionably preserve and provide access to the journals' content, as here defined, over very long periods of time (i.e., 100+ years). Users of the archive must feel confident they are seeing the data and discourse the author and publisher agreed to publish, and not some other version of the work (e.g., a pre-print) or versions of the work corrupted by non-authorial intrusions. By contrast, we hypothesize there is no comparably important idea of archival functionality. Authors, readers, and publishers alike expect functionality to improve over time and independently of any specific content. There is rarely a compelling interest in maintaining earlier functionality (except where there is a failure to improve, as for instance when changes in functionality threatened access to content). And while it is certainly possible to imagine archival inquiries that turn on the functionality provided by the publisher at a given point in time, one might well wish to archive functionality separately and in a manner different from content.

Similarly, because archival inquiries are unlike those that shape the business practices of the publisher, the functionality offered by the publisher may not be what would best support archival use. Indeed, if one accepts that much archival use is unpredictable and requires the repurposing of archival materials, any idea of a single "authentic" functionality may be a barrier to effective archival inquiry.

4, TECHNICAL MEANS FOR DISTINGUISHING BETWEEN CONTENT AND FUNCTIONALITY. Technically, how might an archive of Elsevier Science journals separate content from functionality and ensure the authority and authenticity of the former for the long term?

In addressing this question, we distinguished among four fundamental representations of content carried in digital bits and bytes.

• Alphabetic, symbolic, and BLOB codes. Digital bits and bytes encode alphabets and symbols, using ASCII and Unicode conventions, and a variety of digital objects, using BLOBs (i.e., binary large objects). BLOBs are used to encode tables, images, multi-media objects, etc. ASCII code is one of the most stable features of the digital environment because it is so basic and flexible, because it is used so pervasively, and because it represents massive sunk-costs. It makes economic sense to focus archival effort at this level of coding.
• Document codes. While more recently developed, SGML and related document codes (representing many paratext elements) are a second fundamental level of coding that is essential to create meaning from digital bits and bytes. We expect document codes to be relatively stable and for that reason to be an appropriate target for archival effort.
• Metadata. Metadata relating to a given document is a particularly important category of encoded information. Standards for such metadata are now being developed, with the intention that such information will remain relatively stable in the otherwise fast-changing digital environment. Metadata will be particularly important to and an appropriate target for our archival effort.
• Rendering software. The library as an archival agency must provide its readers with the ability to render alphabetic, symbolic, BLOB, and document codes into a form that can be read and used. Today, a number of choices present themselves, including OpenText, Science Server, and IBM's Digital Library. Other choices will no doubt present themselves in the future. It is possible that some rendering software will be more effective in supporting a particular line of archival inquiry than others. Especially given the unpredictability of archival inquiries, it will probably be important that the alphabetic, symbolic, BLOB, and document codes be renderable by a number of different software packages.

5. CREATING A BASIC ARCHIVE. Elsevier Science is able and willing to provide the Yale Library with SGML versions of nearly all of its current journals (c. 1,100 titles). The availability of back files in SGML form varies, with significantly less SGML material being available for pre-1998 publications. Other digital formats-including TIFF and PDF-wrapped TIFF files-are available for pre-1998 material. Elsevier Science does not wish to provide and the Yale Library does not wish to use for either of the two services described here the full online functionality offered to Elsevier customers. Elsevier Science will, however, provide to the archive all of the metadata it creates. In its simplest conceptual terms, therefore, our agreement is that Elsevier Science will provide the content for the archive (i.e., the bits and bytes coded for rendering alphabetic and other symbols, BLOBS, documents, and the metadata associated with this content), while the library will provide rendering software and the computing environment for its use.

In creating that environment, we will draw on the inherent advantages for archiving of a multi-tired client/server architecture supported by a mass storage archival system. Tier three of the system will provide database management functionality where the content will be stored in a RDBMS, ODBMS, or a file system. Tier two, our middleware, will provide process management software. We are drawn to the archival management software being developed at the San Diego Super Computer Center and at the Harvard and MIT libraries (our partners in the Northeast Library Consortium, or NERL) to manage the intake, object naming schemes, validation, access management, and other tasks that must be performed at this tier. The design of the these tiers will be informed by the InterPARES Project, a leading effort to develop the knowledge required for the permanent preservation of digital records, and by the standards for digital archives being advanced by the Council on Library and Information Resources and the Digital Library Federation. Tier one of our archival system will provide a user interface, in the form of a web server-browser, to request, render, and display data from tiers two and three.

The content and computing architecture just described will constitute a digital archive capable of meeting the two most essential archival responsibilities. These are (1) preserving the documentary content published by Elsevier Science and the evidence of provenance relating to that content, and (2) supporting a wide range of largely unpredictable inquiries against that content and its provenance.

At the same time, we intend to design our archive to be technically capable of meeting the ordinary, non-archival information needs of readers should (or when) commercial circumstances prompt Elsevier Science to change its business practices in ways that imperil access to its content. The Yale Library and Elsevier Science will identify a set of specific circumstances in which the library's archive will become the means of ordinary access to Elsevier Science journals. These will include the business changes identified above, such as when parts of the publisher's content becomes commercially non-viable, when publication of a title ceases, when the publisher substantially changes its line of business, etc. The Yale Library and Elsevier Science will establish agreements governing the intellectual property affected by changed circumstances such as these. We will also describe the computing environment that the Yale Library will provide to support ongoing access and the latitude the library will have for non-profit business arrangements that might improve the provision of access. Both collaborators will publicize their agreement on these matters, to inform authors and the wider library community about the steps we are taking to ensure that enduring, long-term access to our archival content is certain.

The Yale Library already has substantial experience in working as trial local-load customer of Elsevier Science. This experience prompts our interest in a preservation project, and we are particularly glad to have the strong support of the Elsevier Board of Directors for this project. Though we will provide an archival home for only one publisher initially, considerable variety in the size and content of Elsevier's publications ensures that our technical design will be generalizable and scalable. Moreover, we believe the example of a successful project with one of the world's most important publishers of science journals will be influential when we approach other publishers, later, and when other libraries and publishers consider collaborative action on archiving. Finally, the archival profession urgently needs some highly visible success stories regarding digital content. The ready engagement in and commitment of Elsevier Science to this project will be important factors in our success in planning and implementing archiving services.

6. A DEVELOPMENTAL ARCHIVE. Although we feel confident about our ability to design a digital archive able to meet basic archival responsibilities and provide against an important set of commercial contingencies, many questions will remain about the good design and management of a digital archive. In our initial discussions, the Yale Library and Elsevier Science have identified the following questions as important to both of us. We expect to refine, refocus, and prioritize these questions and to identify others during the planning phase of our project.

Questions bearing on the technical design of the archive

• How persistent will the coding conventions be for alphabets, symbols, BLOBs, documents, and metadata? Are there real economic advantages in making these canonical codes the focus and basic building blocks of our archival work?
• A significant part of what Elsevier Science has published digitally exists now only as TIFF and PDF-wrapped TIFF files, and not in SGML. Should this material be included in the archive, even though the more basic coding is not available? This question points to a yet more fundamental one. What do we regard as the most reliable document for archival purposes? Is it a particular rendering of the document (a TIFF file, for instance) or is it the underlying content coding that can be rendered in a number of different ways?
• What are the limits of our ability to separate content from functionality? These two are so closely linked in much software design that we will surely encounter areas in which the separation cannot be made without some jeopardy to the integrity and utility of the archival record.
• What should the archival record of functionality be and where and how should it be maintained? In what measure is the provenance of functionality more closely bound to the publisher's corporate archive than to the content itself?
• What options for data rendering and other middleware software do we have? Do any of these options meet archival purposes better? Do any promise to be more robust for those cases where the archive must substitute for the normal, commercial means of access?
• What resonance does our proposed separation of content and functionality have with thinking within the digital preservation community about migration and emulation strategies? Does either strategy become decidedly more or less attractive given this separation?
• How does one test the reliability over time of the archive as a means of reader access to its digital content?
• Significant differences in media stability and storage capacity suggest there may be a significant choice to be made between tape or compact disk technology for our storage medium. How do these trade-offs bear on our ability to contain the cost of media migrations over long periods of time?
• Our focus on the fundamental alphabetic, symbolic, BLOB, and document codes assumes relative stability in those codes, some of which are just now being developed. It may be that over very long periods (100+ years), even these basic codes will change so much that emulating them or migrating through their changes will be formidably expensive, especially in light of presumably infrequent use of archival material. Given this possibility, should we explore a technically bi-modal archiving strategy that preserves the content (not the functionality) as digital output to microfilm, which could be reconverted to digital format on demand where that re-conversion was appropriate for a particular archival inquiry? If there is merit in such a bi-modal approach, how would certain information (e.g., color images, video and audio material) be handled?

Questions relating to changes in commercial circumstances

• Can we anticipate some kinds of commercial change and manage them proactively? For instance, can we identify a set of ordinary uses of Elsevier Science content that can or should be segmented because they pose distinctive commercial challenges? Should we actively plan for different commercial strategies to maintain access to current materials (including, say, material published in the last 5 years), recent materials (e.g., material published 6 to 15 years ago), and older materials (e.g., published more than 15 years ago)? At some point, access to infrequently used material will not be profitable and will burden the commercially viable parts of the publisher's business. How can we measure and understand such information life cycles in a way that serves both the commercial needs of Elsevier Science and the access needs of readers, where those needs diverge? Are there technical or other operational synergies among a publisher's profitable services, its emergency back-up services, and the not-for-profit operations of an archive that can be invoked to manage through this information life cycle?
• What resources-human, technical, and financial-will be required to sustain the archive indefinitely over time? What contribution will Elsevier Science make, as the creator of the archival material, toward meeting these resource needs? What contribution will the Yale Library make as an archival agency? What contribution would other libraries, which depend on the existence of the archive, make? Are there points in the life cycle of Elsevier Science publications where marginal revenues might be realized but might not justify the commercial effort required to earn them? If so, might these revenues be used to help support the cost of archival operations? (A rough analog is the way some university presses are now using mid-list books that are no longer sufficiently profitable to attract trade publishers to build income to support traditional scholarly publishing.)

Questions about the utility of the archive

• What view will authors take of the proposition that the digital coding of their documents, rather than a particular rendering of those codes, constitutes the archival record of their work? More generally, what view will authors take of the systematic differentiation between content and functionality that we propose? These same questions need to be posed to users of the archive as well.
• When commercial changes shift readers away from the commercial services of Elsevier Science to the archive, in accord with our agreements, will a technical structure appropriate for infrequent archival inquiries adequately support possibly more frequent ordinary use? In such circumstances, the archives middleware and telecommunications capabilities may not be appropriately scaled and they may not offer some functionality important to readers. How will these issues be addressed, and what not-for-profit business arrangements might be instituted to solve problems of scale that arise in such circumstances?
• What kinds of inquiries will be made of the archival record we create? Are we right in predicting they will be sufficiently infrequent, sui generis, and involve such transformative uses of the record that they pose no attractive commercial opportunities for Elsevier Science? Can the library maintain a fully open (or "bright") archive that responds usefully to archival inquiries but does not seriously threaten the commercial interests of Elsevier Science? More broadly, what demand for related archival resources will an archive of journal content create? What impact will our content archive have on broader records management and archival practices within Elsevier Science?
• How do the costs of maintaining a digital archive of Elsevier Science publications compare with those of maintaining a print version of the journals to ensure long-term preservation and access? In considering this cost differential, how should we appraise the capabilities of digital publishing that cannot be replicated in print?

Questions that may affect the strategic direction of Elsevier Science

• Given the likely difficulty in some cases of disengaging content from functionality, are there things that Elsevier Science might do in the design of its current and ongoing business process to facilitate that separation? That is, in what ways can its commitment to good archival practice become a source of strategic direction for Elsevier Science?
• The metadata provided by Elsevier Science will be the primary testimony to the provenance of the documents in our archive. How adequately do these metadata represent the processes by which Elsevier Science conducts its business and creates a public (i.e., a published) record of scientific communication? Are additional metadata required? Here again, in what ways might its commitment to good archival practice become a source of strategic direction for Elsevier Science?
• Is there provenance information about the documents in our archive that might best be maintained in a separate corporate archive? For instance, is information about editorial board membership for a given journal best kept in our archive of individual published documents or in a corporate archive of Elsevier Science? Where should the archival record of the functionality Elsevier Science created and delivered to its customers be kept, in both the near and long term?

• Acting alone, the best that any publisher can do to ensure enduring long-term access to its content is to make credible arrangements to transfer that content to a library, should commercial circumstances make that necessary. This is true of Elsevier Science as a commercial publisher, just as it is true of JSTOR, the nearest we have today to a trusted third-party agent for preservation and access. There are, however, no agencies to which libraries can hand off their preservation and access commitments, should they need to. Is there anything that can or should be done about this fact? Is there anything about this fact that has changed with the emergence of digital technology? Is there anything that can or should be done now to protect the preservation mission of libraries against the pressures created by other parts of their mission?

7. INITIAL PLANNING ACTIVITIES. Few of the questions posed above can be usefully explored-either by the Yale Library and Elsevier Science or by other investigators-unless we actually create a digital archive. There follows an outline of the planning steps we will take over a year's time to create a digital archive for the journals published electronically by Elsevier Science. An italicized statement at the end of each bullet indicates the sequence and duration of our planning activities.

• Decide (for at least initial implementation) on the content of the archive. Will it include TIFF and PDF-wrapped TIFF files as well as SGML files? Will it include a variety of editorial matter (names of editors, instructions to authors, copyright policies, etc.) as well as content created by authors? Will the archive include advertisements appearing in the print versions of the journal? Does the archive require any metadata beyond what Elsevier Science now routinely creates? Will the archive include ancillary indexing and abstracting services or free information resources now created by Elsevier Science? The decisions we make on these issues need not be regarded as definitive or final. Indeed, we may wish to make different decisions on some of these matters for various parts of our archival file, so as to test the utility of our initial answer to these questions. First quarter.
• Establish initial agreements regarding intellectual property and other business practices that will govern the archive as the means of supporting ordinary use of Elsevier Science content when changes in commercial circumstances make that necessary. First quarter.
• Acquire and deploy the test bed computing hardware and software needed for the archive of Elsevier Science publications described above. Explore the willingness of software developers to provide their products to us for short-term testing and development activities at no cost or at below-market rates. During the planning year, only a test-bed level of capability will actually be deployed. Production level capabilities would be created in the follow-on demonstration project. Nonetheless, planning would attend carefully to questions of operational efficiency and to questions of scale as regards the initial size of the archive, its growth over time, and the requirement that the archive be capable of supporting some approximation of normal use should commercial changes require the archive to support such uses. Decide whether the archive will support more than one capability for rendering the encoded material of the archive. First and second quarters.
• Devise and implement prototype procedures to ensure both the reliability of the archive's content (i.e., it faithfully represents the editorial and publishing activities of Elsevier Science) and its authenticity (i.e., material in the archives persists without change, except such changes initiated and documented by the archive as part of carrying the material forward through time). These activities and those in the next bullet will require substantial travel and consultation, to ensure that our project takes full advantage of the archival work being done at other NERL partner libraries and by R&D laboratories elsewhere. First and second quarters.
• Devise and implement prototype methods by which the operational stability and reliability of the archive can be tested periodically. The object here is to create a solid record of access, for both archival and normal inquiries, that leaves the community served by the archive confident of its technical reliability. Second quarter.
• Devise and implement prototype methods to register and assist readers pursuing archival inquiries, comparable (where appropriate) to those used in print archives. Devise and implement service quality improvement procedures that use our experience with readers to improve the archive. Third and fourth quarters.
• Develop a fully-detailed technical, service, and business plan for the creation of an enduring archive of Elsevier Science publications. Design this plan so that it might accommodate the journals of other publishers. Third and fourth quarters.
• Continually revisit, refine, and refocus the questions posed in the previous section. Restate them as appropriate and begin to devise the methodologies that may be used for addressing them, giving special attention to those that bear on the long-term sustainability of preservation and archival functions. All four quarters.
• Prepare project announcements and reports, as appropriate, for the Mellon Foundation; for the publishing, library, and digital preservation communities; and for authors writing for Elsevier Science publications.

8. PROJECT STAFFING AND BUDGET.

Staffing. Scott Bennett (Yale University Librarian) and Ann Okerson (Associate University Librarian) will co-direct the project, drawing on their substantial experience in scholarly publishing, preservation, copyright, and licensing. They each expect to devote not less than 5% of their annual effort to the project. They will in fact devote whatever time is required to ensure the success of the project.

Karen Hunter (Senior Vice President, Elsevier Science) will direct the Elsevier Science part of the project planning effort, drawing on her distinguished record in scholarly publishing and on her career-long liaison work with academic libraries.

Paul Conway (Head of the Yale Library Preservation Department) will be the project manager. He will devote an estimated 40% of his annual effort to the project, drawing on his internationally recognized expertise in preservation and digital technology and on his demonstrated project management skills. Working under the direction of Scott Bennett and Ann Okerson, Paul Conway will keep the project team focused on the completion of its tasks in a timely fashion. He will provide the project with daily coordination, leadership energy, and staff support; he will draft most of the project documents.

The project will hire a systems officer to lead in specifying the technical environment needed for our archival work, to explore alternatives for creating that environment, and to build prototype elements of our system. We expect to employ a systems officer already working at Yale, to ensure familiarity with the institution and a prompt start for our project. Practically speaking, it will probably not be possible to release the full-time effort of a senior systems officer for this project. Preliminary conversations suggest we can secure the majority of such a person's time, along with the efforts of others, to give us a strong team with a wide range of expertise and the equivalent of at least a full-time appointment.

Project personnel have substantial experience in defining and negotiating licenses for intellectual property. Yale's Office of the General Counsel has been both thoughtful and generous in supporting this work when that has been needed. We believe this tested team approach will meet the project's need for legal counsel during the planning year.

While it is likely the project will need the services of a business planning consultant, the nature, scope, and cost of those services cannot realistically be predicted now. It seems best to rely on the University Librarian's Discretionary Funds for this purpose.

return to top >>