KMWorld CRM Media Streaming Media Faulkner Speech Technology Unisphere/DBTA ITIResearch.com
PRIVACY/COOKIES POLICY
Other ITI Websites
American Library Directory Boardwalk Empire Database Trends and Applications DestinationCRM EContentMag Faulkner Information Services Fulltext Sources Online InfoToday Europe Internet@Schools Intranets Today ITIResearch.com KMWorld Library Resource Literary Market Place OnlineVideo.net Plexus Publishing Smart Customer Service Speech Technology Streaming Media Streaming Media Europe Streaming Media Producer



Magazines > ONLINE > May/June 2010
Back Index Forward
 




SUBSCRIBE NOW!
Online Magazine

Vol. 34 No. 3 — May/June 2010

FEATURE
Ebooks: From Institutional to Consortial Considerations
By David Stern

While many academic libraries are still exploring the utility and acceptance of ebooks, either as replacements or supplements to paper monographs, it is not too early to begin considering the implementation issues for consortial purchases and/or leases of ebooks. These considerations add new layers and entirely new elements to the already complex ebook models still under exploration by individual libraries.

While electronic books have been available for a number of years, there is still little agreement about best practices, with many possible variations that make selection, payment, cataloging, searching, and presentation extremely complex.

Additional layers of options and decisions arise when consortial considerations are added to the scenario. Few of these new issues have been explored or tested by large consortia. A number of small consortial arrangements are being tested and will provide some guidance, but large consortial approaches involve very different considerations in terms of standard pricing models, cooperative profiling, and shared payment methods.

SINGLE ORGANIZATION EBOOK CHOICES

Before moving to consortia considerations, what single-site considerations are relevant?

Will the ebooks actually be used? Which portions of the population will adopt them first? Will acceptance be based upon discipline characteristics or age-based behaviors? Will the ebooks be used as simply distributed versions of paper books, or will new functionality be offered and adopted? What new options are desired and required—better screen displays and advanced functionality, such as full-text search, enhanced navigation, and personal annotations? Will most readers prefer one single platform or multiple branded versions on publisher or host platforms? Will these ebooks be supplements or replacements for paper materials? How do we deal with multiple purchases and reserve copies? What cost models and ordering tools are acceptable?

While many uncertainties and unknowns remain, some options appear to be finding traction with readers, libraries, host platforms, and publishers, which may lead to best practices. In other cases, library experiences have only raised additional questions; there is no agreement or direction pointing to the obvious choice for most types of libraries. Current implementations are somewhat chaotic and time-intensive, as the competing players continue to develop and test new models. The good news is that many hosts and publishers are now involving advisory councils of librarians in their explorations.

APPLYING SINGLE-SITE LEARNINGS TO CONSORTIA

With materials across all subjects, the Academic Complete basic ebook package from ebrary shows adoption for most disciplines, despite anecdotal comments from subject librarians stating that their populations are not interested or ready for such tools. Those performing initial reviews of use data find the alignment between user demographics and collection use. Subject discipline and age are not driving factors for modern researchers. Additional studies need to determine whether these initial ebook uses were adequate replacements for, or merely supplements to, paper copies.

Anecdotal evidence shows that early ebook adopters appreciate the added value provided by ebooks. The power and functionality offered by enhanced markup using XML (as opposed to simple PDF presentation) allows for novel search and navigation possibilities and has provided interactive tools containing options such as the manipulation of equations and graphs, as seen in the Knovel platform. This same markup could also be used to allow for the purchase of parts of ebooks once the unbundling of materials reaches maturity.

MYSTERIES OF SEARCH

Searching remains a complex and mysterious operation for users. In much the same way that multiple journal platforms create incomplete searches against selective branded search interfaces, the many ebook platform options also create confusion for users. The wide variety of interfaces for searching parts of the available ebook material raises questions of whether you are searching reference materials or trade books. The average researcher expects to find all types of materials within the OPAC, but this is rarely the case. Most OPACs do not even search the complete full text of ebooks, limiting search to MARC information.

The use of federated searching tools should provide the full-text searching that is possible. Eventually, these tools may be replaced by harvested and enhanced index searchability. Yet questions remain: Do we offer searching across all available ebooks or limit searching to those ebooks that the library has already purchased? Speaking of seamless searching across all domains, should we create search tools that automatically search deep into full-text books from free platforms such as Google, the Internet Archive, and Project Gutenberg? Consider user preferences: Once an item is located, will our ebook platforms support all reader devices, such as the Kindle, the iPhone, and smartphone and cell phone standards?

MANAGEMENT AND ACCESS MODELS

Assuming users do not want to learn multiple interfaces for searching ebooks, they probably also would like to have a standard search and citation tool (RefWorks or EndNote). Even better, they will probably want a knowledge management tool (Zotero) that can handle all types of information sources and formats (books, journal articles, PowerPoint presentations, and other web sources). In some cases, libraries will still need to offer platform-specific tools with special features (chemical structure builders).

Will these access permissions continue to be based upon ownership models, or will we incorporate larger elements of lease and/or pay-on-demand models? What about the viability and scalability of tracking annual use data for all leased material in order to make continuation decisions? Some price models with platform fees are based upon a number of active ebook subscriptions. If these approaches are to continue, they will need to offer better methods of use data analysis over time periods for making informed decisions.

With title-level orders, subject-based packages, and cross-publisher packages comes a concern about duplicate titles. This is made even more complicated by the inconsistent handling of titles within sets as individual purchases and within various packages. Approval plan blocks have not yet been implemented across many ebook approval plan profiles. Examples of overlap exist in many of the big packages—ebrary Academic Complete, Knovel, Elsevier’s MD Consult, McGraw-Hill’s AccessMedicine, ProQuest’s EEBO, and Past Masters to name just a few.

Updates to ebook packages create labor-intensive situations for libraries attempting to remain accurate and current. Selectors must review books added or removed from packages, consider possible alternatives, and update systems. Some commercial products can update MARC records for the OPACs and even update resolvers for ebook links, but what about all the users who come through Google and Amazon and don’t utilize the virtual private network (VPN) for redirection to the accurate information? Few libraries proactively check for accurate links, especially for individual titles within big packages. Reactive corrections made only after users complain provide very bad service and will hinder the adoption of ebooks.

Too often, platform differences guide our choices. Some platforms offer full downloading but only very limited page printing; other platforms have more generous page print limits (albeit with frustrating workarounds included) but don’t offer downloading. Some platforms do not address price options for single users versus multiusers, nor do they offer short-term reserve use options.

Internal processes within individual libraries

The majority of new policies and procedures for handling ebooks are not overwhelmingly technological in nature. But they do require new ways of interfacing among libraries, readers, and vendors, as well as across related support services. Examples of current methods that require revision include the following:

How we order: Ebooks allow for immediate identification, purchase, and access by end users. Complexities include seamlessly allowing for ordering ebooks from our standard vendors, ordering direct from specific publishers, and addressing license review and access options such as multiple seats or platform preference. Libraries need to develop templates for ebook orders that include both technical considerations and procedural steps. Not every selector will be able to answer all the questions, so acquisitions support may be necessary for many requests. If desirable, a library must block paper ordering and update MARC records after patrons order ebooks, or at least until a decision is made by a selector if patrons only send online recommendations to librarians. If there is a price break for ordering both the paper and the online versions, that would require better and different business relationships between ebook and paper book vendors, as not all have seamless reciprocal purchasing options—in some cases, these may directly conflict with their existing business models.

How we pay: Libraries traditionally paid for books as one-time purchases or as parts of standing order sets. There were no platform fees, aggregator fees, third-party support costs, or annual maintenance fees. In many systems, there is no way to show and account for such continuation commitments. There are also few options to handle and reflect multiple seat charge options, revised charges based upon use data, reserves use fees (as a short-term option or as a permanent right), or storage of use data and continuation decision histories.

How we review: Most libraries do not perform annual analysis on previously purchased books. There may be use analysis for sending materials to off-site shelving, but that is different from making continuation decisions in relation to annual fees. We also need to perform URL checks to be sure we have access to ordered materials that are virtual in nature; this may be a logical place for vendor service or coordinated efforts for economies of scale. We will need to review large amounts of use data in order to make informed decisions based upon actual data and use patterns rather than relying upon anecdotal opinions. We will also need to incorporate user evaluation methods, performing surveys and utilizing focus groups in order to obtain a more complete picture of user reactions. We will also need to revise our reviews of interlibrary loan (ILL) requests.

POSSIBLE IMPACTS OF SIMPLE ADOPTION

The adoption of new policies and procedures will have intended and unintended consequences for the library and library users. Some policies may be seen as positive and empowering, but others may be considered an added burden by some users. The library must promote these new options in ways that demonstrate the efficiency and effectiveness for our users. The emphasis for faculty members should be on how this will better allocate their discipline dollars to their specific areas of interest. The new steps for selectors and support staff should be presented as a way to customize services to user preferences.

There is no doubt that some models of user ordering involve a higher level of unpredictable and unknown spending. This concern can be addressed by limiting the domain and factors allowable in user-based ordering systems. Providing notifications of possible titles and limiting user ordering to selected populations (with URLs, passwords, or Shibboleth) will help contain costs.

At the 2009 Charleston Conference (www.katina.info/d/2009presentations), a number of reports on initial findings from the analysis of ebook initiatives indicate that coverage, usage, and expenses are not significantly different when users order materials compared to librarians selecting materials. In some cases, the user selections appeared to be far more appropriate, measured by uses per title and comparisons to demographic spreads. Savings were possible on a statewide basis by eliminating duplicate paper copies for middle-use titles when ebooks could be delivered on demand for those extra libraries.

What about duplication? Will ebooks be considered replacements for, or serve as supplements for, paper books? Will an ebook mean that ILL of paper books is not allowed? Will the library provide print-on-demand support if the paper version is required or desired? Can these rules be subject-specific based upon discipline needs? Will libraries remove the purchase option if a free copy can be found at Google Books, Internet Archive, Project Gutenberg, or other platforms?

Consortial considerations 

What can consortia learn from the changes in individual libraries’ policies and procedures? For consortial ebook model options, additional factors include technical investments and philosophical decisions about the best use of shared resources. As in any group activity, some compromise will be required to obtain the maximum benefit for the largest group. Assuming reasonable flexibility within ebook platforms and models, the degree of similar functional desires may determine the formation and memberships of new consortial groups.

Publishers have arrangements with book vendors for volume discounts, and many traditional paper models are based upon historical multiple copy sales of paper titles through approval plans. New pricing models for ebooks can incorporate similar volume discounts for consortial purchases. However, one of the advantages of ebooks is that there is no longer a need for “just in case” multiple copies. Purchasing can be based upon actual use, or a pricing model could be based upon either historical spends by Library of Congress (LC) range or some new value point—say two and a half times the price of the paper book for a consortium of six libraries with an average overlap of two and a half copies. Publishers desire guaranteed revenue, and many libraries prefer known costs, so there is a benefit to some type of advance pricing for those libraries that do not want to rely on uncertain and variable use driving purchase fees. Some small consortia have started to develop price-point models across multiple publishers, and these tests will perhaps inform and drive larger consortial models.

Vendor value comes from offering economies of scale for both shared book selection profiles and for providing shared ebook platforms, which makes local hosting unnecessary for many materials. This service involves negotiation and contracting for acceptable services with libraries and content producers and hosts. It also requires transparency across organizations that have traditionally guarded tightly their proprietary knowledge and content.

CONSORTIAL PROFILES

Consortial arrangements will require complicated and complex matrices for services such as local selection preferences, subject-based fund assignments, and pricing instructions. Vendors are best positioned to develop industrywide options rather than handling publisher-by-publisher solutions. Selection and discovery assistance will need to be generalized across the consortium but also tailored to local needs, with separate feeds to librarians and users. These consortial profiles will also need to incorporate library-specific rules for issues such as paper or ebook equivalence and duplication. Some libraries may request that orders be passed through to different vendors based upon language, country of origin, or other factors. Other important added-value services by vendors include the ability to handle licensing and library-specific selection preferences, update notifications to companies that supply related MARC records, and updated notifications to other related service providers such as link resolver data services.

Libraries have many decisions to consider and may choose their consortial partners based upon a variety of special interests. General ebook packages may be obtained through one type of consortium, while subject-specific packages and profiles may exist with other consortia utilizing different vendors. Title-specific ordering may be accomplished through yet another consortial arrangement. The ability to coordinate these efforts to avoid duplication and provide one interface for search and ordering is important for efficient and effective operation, especially if users are expected to be involved in the selection process.

MULTIPLE CONSORTIA

Libraries may choose to join consortia to obtain simple multiple-site discounts on their own isolated orders, they might opt for shared costs and discounts based upon overall consortial demographics, or they may choose to pool funds and create a variety of allocation plans using assigned LC ranges or other variables such as historical purchase patterns. Some libraries may decide to return any realized savings to the consortium for broader depth of coverage, while others may prefer to repurpose the reduced spend into other areas such as staffing or service charges. Consortia may select to handle print on demand centrally or on a site-specific basis. If handled centrally, questions arise: To which library will the paper book be returned? How will the added fees be charged? Some consortia may even elect to share an Espresso printer instead of outsourcing the paper generation to a third party.

Will consortia elect to remove the purchase option for paper after an ebook is ordered—or perhaps send only recommendations to librarians in this case? Will these decisions be centralized or will members be able to make local decisions within the consortial framework? In a similar way, will the consortia elect to remove the purchase option if a viable copy is found at Google Books, the Internet Archive, Project Gutenberg, or other platforms? Will this be universal or will it be the policy for only some libraries within a consortium? For space-savings, will the consortium agree to replace and/or reduce older paper copies of little-used titles with on-demand ebook purchases or accept a Google Book option as an alternative? Finally, many of these decisions are made in relation to local licenses and legal requirements—which differ significantly by state and by library type.

SEARCHING AND METHODS OF ORDERING

The following are examples of preliminary considerations realized during an initial planning effort by a subgroup of libraries within a large research library consortium. Many of these same questions will be relevant for consortia of all sizes as the matter of scale does not affect the underlying concerns. Early conversations with major vendors helped in developing these understandings, and planning is now underway within the vendor community to provide practical options for such initiatives. It is still too early to provide specific solutions, but it is now time to have additional librarians with a variety of interests involved in the conversation.

Types of Ordering

The first major issue considered was the method of identifying and ordering ebooks across the consortium. We needed to understand the possibilities, practicalities, and implications of supplementing our traditional librarian-selection model with user selection. A number of levels of user selection options ranged from placing MARC records for possible ebooks into the OPAC through providing seamless full-text searching within all available ebooks as part of a standard book search.

For patron order-on-demand services, the intention was for users to perform basic book searches federated across all platforms (local catalogs and the entire ebook platform contents, including full-text searching and their associated MARC subject headings). In this scenario, it would not be necessary to load MARC records into local catalogs and then remove them if ebooks were not purchased over a certain time period. This would provide far better recall with little local effort.

If libraries were concerned about the possible costs of providing books from across the entire possible ebook domain, a library could limit searching to only certain preselected subject areas. Another cost containment option is to tailor search domains and available services for factors such as home location or patron type. What about a suite of patron-order-on-demand controls in order to provide the most effective access to specific discipline materials based upon annual resource allocations?

We explored push methodology to send electronic notifications to selected faculty members, which allows them to place recommendations for newly released paper and/or ebooks into a librarian’s ordering environment. This provides immediate and intimate contact with faculty in book selection. Encourage faculty members’ involvement in specific subject areas by providing special URL access to let them search the full text of all possible ebooks and obtain immediate access through automatic ordering.

The ebrary Title Preview service (http://site.ebrary.com/lib/librarytitles/home.action) already allows for this deep searching across all metadata and full text. For actual selection and ordering implementations, the domain can be restricted by LC range or keywords. Shibboleth overlays would allow for restrictions by patron type if desirable. For selector purchasing, approval plans are being migrated to cover ebooks, but decision factors must include new variables such as preferred platforms, preferred number of seats, and purchase or lease and include or exclude paper copies if published within a designated time period. These decisions may be universal for all consortium members, or there may be opt-out or special considerations for selected members (local author coverage, media types, etc.). It may be possible to develop an administrative module to select specific ebook titles per library using a variety of subject and nonsubject parameters, but the vendors would prefer to have one profile for the entire consortium.

Another approach would be to have a profile-generated review file and a designated review period with any ebooks not removed from the review file charged to the consortia, as with a paper approval plan. Consortial members would need to determine which selectors review which files, and files may be generated for varying expertise levels within any subject area. For less important areas, notification slips could also be used to provide additional options for coordinated selector purchasing. Cooperative paper books collection development has been performed by libraries using shared approval review systems for some time. This ebook review process would allow for even more specific and immediate decisions utilizing shared funding or cooperative understandings.

Types of Payment

Payment models for individual libraries already allow for complex funding options, but consortial funding allocations and reporting will be far more complicated and will require central tracking for annual analysis across the group. Facets may include languages, media types, platform-specific media versions, cataloging fees, service charges, multiple copy fees, paper copy fees, and reserves fees. Fees may be divided equally across members or may be determined outside the vendor system with a central consortial payment made to the vendor. Or some type of real-time charging mechanism may exist with tracking and protections added for unusual spending situations. In some cases, a predetermined annual payment may allow unlimited access to packages of material. Permissions for users will need to be determined based upon recorded VPN location verification and/or Shibboleth validation.

Consortia will need to develop either shared allocation equations or profiles for sharing expenses. These may be single rule systems or they may provide for a variety of levels of contributions based upon factors such as population size, discipline strengths, languages covered, and subgroup members.

Some consortial arrangements may work best in completely separate consortial partnerships based upon shared collection missions and resources—imagine a special Divinity consortium outside of the normal consortium used by the regional libraries for graduate-level English and political science materials. Or imagine a national consortium for the new field of nanotechnology materials; why restrict members to regional libraries if the group is obtaining easily distributed online materials? Consortia may become larger and less geographically based; witness the SCOAP3 high-energy physics initiative, which is an attempt to become a global consortium for journal materials, but there is no reason this could not also work for ebooks.

SUBJECT AND EXTENDED OPTIONS

Consortia would need to bill individual libraries for percentages of purchases or allocate subject areas to various libraries based upon equations or profiles. Some reciprocal subject purchasing could be arranged among libraries collecting at similar levels. Present consortia could utilize existing buying patterns, as determined by statistical analysis of current paper book purchases, to start these calculations. Baker & Taylor’s YBP’s (www.ybp.com) collected buying-patterns data for paper books can be used within subject classifications to determine appropriate costing metrics.

Libraries would then need to decide and enter into their profiles rules for additional charges. For instance, would charging be directed to the requesting library or be shared by all (or the designated subject library) if paper copies were supported? Where would the paper copy be housed after use? Perhaps in the library designated to cover that LC subject area if there were such an arrangement?

The system might also be set to charge users via credit card for on-demand printing for their own use, at least for libraries willing to offer such a service. In some cases, such as faculty requests, the rules might offer user-based ordering, perhaps with some sort of librarian oversight in the background. This option could also be controlled by a profile preventing overexpenditures and limiting the LC ranges that contain this option. This becomes a complex set of rules overlayed onto a complex set of initial subject profiles. Accurate and current verification and validation would be quite important for satisfactory user experiences.

CORE COLLECTION SUPPORT

Initial allocations could be handled as equal “general fees” for core collections based upon historical spends against established subject profiles. Libraries could then contribute reduced fees for materials in lesser-used areas, and selected libraries could provide additional contributions for more advanced collection building in designated research areas.

These initial payments may be handled internally by the consortium or they may rely on ebook vendors to arrange for at least portions of the charging mechanisms—particularly when looking at patron-on-demand options.

The vendors are interested in working with coordinated libraries in order to convince publishers that these consortial pricing models can produce additional byproduct revenue from after-discovery paper book purchasing. Ebook discovery has certainly led to higher paper ILL book requests.

Libraries will need to more carefully monitor actual ebook use and preferences for online resources—particularly when the statistics are eventually provided in a meaningful way and there is a critical mass of material on which to observe normal behaviors in specific domains.

LONG-TERM STEWARDSHIP CONCERNS

Consortial collaborative collection development provides solutions to address the issues of monograph redundancy and the effectiveness of ebooks to reduce the footprint required to satisfy user needs for local copies. Questions about the optimum number of book copies required for sustainability have traditionally discussed dark archives of paper copies and multiple copies of circulating copies. Ebooks can provide on-demand access and quick printing, reducing the number of required copies; the consortial savings from such actions could be enormous and could subsidize additional digitization efforts to expand the materials covered by this approach.

Preservation of digital and born-digital materials is still a topic of debate. Consortia are the logical groups to explore these details and to develop best practices to ensure safe storage, functionality, and access. The associated question of migration to new formats and/or emulation needs to be discussed during this review.

Integration and scalability becomes another group concern as enhancement to ebooks, such as personalization and annotation, become the expected ways of interacting with digital materials. The level of shared research notes and social networking comments is an infrastructure issue that needs to be incorporated into these early design phases of ebook systems. The handling of personal links raises the issue of levels of security when dealing with group work, historical materials, associated raw data files links, and research privacy concerns. The knowledge management aspects of shared networks is a topic for another article, but they must be addressed before too much design work begins in these shared resource networks.

If we intend to have a logical set of options, both at the single institution and the consortial level, librarians must influence the industry through proposing best practices and offering alternatives. The top priority for the future is to have a continuum of options from locally customized plans to easily understood purchasing profiles based upon standard criteria. Ebooks are fast becoming standard fare. Libraries should move quickly to develop policies and procedures that allow ready incorporation of them into collections.

Action Items and Considerations for Cooperative Collection Development of Ebooks

I. Immediate Possibilities

A. Reduced costs for multiple copies, with librarian selection

Without making any serious commitments, it is now possible to create a YBP ebook account that allows shared viewing of purchasing activities across designated profiles/libraries. This account could start with slip notification for librarian selection and coordination of available ebook options, and no actual automatic selection and delivery would be initiated. The ebrary platform is very interested in exploring and arranging reduced pricing for multiple “copies/seats.” This reduced pricing is now in place for a small cooperative collection development group in Maine.

B. Searching/ordering across all titles within the ebrary platform

Again, without any commitment, current YBP GOBI (Global Online Bibliographic Information) searching of ebook metadata could be supplemented by librarians performing full ebrary searches (MARC, full text, other). Integration efforts under consideration should allow for YBP orders using ebrary search results in order to provide MARC records and track payments.

Required Action:

For libraries, the only requirement to begin these initiatives is to develop a shared YBP account for consortium members. The membership could develop shared profiles to reduce redundant review efforts. The selection of individual items would be on the shared plan in order to coordinate actions, identify duplicate orders, and determine the multiple copy cost reductions.

The ebook vendor, in this case YBP, would create one shared ebook account for the group and use its GoBetween software to allow consortial viewing. Profiles could be created by LC (Library of Congress) range and publishers in order to produce customized new-book notifications.

The ebook platform, in this case ebrary, would negotiate reduced costs and price points with publishers for selected consortia (library groups).

While these possibilities technically exist, there are some logistics that need to be addressed to make this system optimal.

First, there needs to be a better industry method to provide timely information about confirmation and/or delay in the release of ebook versions compared to paper versions. Second, there needs to be a clear statement of intention and notification for ebooks to be included in aggregator platforms (e.g., Academic Complete) in order to avoid or limit redundant purchasing. Third, there needs to be a policy for assigning the cost reduction across sharing libraries: We don’t want all libraries to wait for the first full-price copy to be purchased before buying the reduced-price second multiple copy.

While the examples listed here involve YBP and ebrary, Coutts is equally well-positioned to provide such service. The coordination of orders across libraries will require either all libraries to utilize the same agents or the development of very sophisticated interfaces to maintain histories and provide real-time group decision making across various services. This issue of scalability is going to be a key element in developing efficient and effective networks that cover the broadest possible domain of ebook materials.

II. Future Initiatives

A. Reduced costs for multiple copies, with patron selection

This enhancement allows researchers to search the entire ebook database (or preselected portions for specific libraries) and to place point-of-need orders. It would be possible to start by having recommendations go to librarians for review—until costs, patterns, and implications are better understood. This model could start with a simple price reduction for multiple copies, or it could immediately include shared costs according to profiles. Providing the database access is not terribly complicated, but customizing and tracking orders will require new processes and procedures. Determining the allocation of funds per library will be the complicated element if this is to include a shared savings arrangement.

B. Tiered costs for multiple copies, with librarian selection and/or patron selection

Based upon prearranged profiles for LC ranges using a fee structure based upon prior buying history, this cost allocation model could adopt the earlier search and order processes. LC range allocations could be consortiawide or library-specific groupings. Designated libraries may agree to cover costs for reciprocal LC ranges to simplify or reduce internal payment transactions. More complex issues would address which libraries would include cost coverage for paper copies and the final locations for paper copies—perhaps to either the LC-designated library or the requesting patron location. These special arrangements would need to be reviewed for appropriate cost sharing over time, especially with changing research interests and likely modifications in user behaviors.

David Stern (hdavid.stern@gmail.com) is an independent consultant who has worked at Brown and Yale universities.

Comments? E-mail the editor. (marydee@xmission.com)

       Back to top