NLS: That All May Read

Current Strategic Business Plan: Implementation of Digital Systems

March 2004

Executive Summary

This summary presents an overview of the current strategic business plan for the implementation of digital systems and services for the free national library program operated by the National Library Service for the Blind and Physically Handicapped, Library of Congress (NLS), its network of cooperating regional and local libraries, and the United States Postal Service (USPS). The program was established in 1931 and is funded annually by Congress. While there are still several important "unknowns" at play that will determine the technological and economic feasibility of implementation for alternative digital systems-unknowns that will be addressed and answered in several future studies-this summary provides an overview of a comprehensive strategic plan to facilitate the conversion of the program to digital technologies. The plan will be updated and refined as supporting future studies are completed.

While analog cassette technology has been the backbone of the program for many years, it is now outdated in several respects and is nearing the end of its useful life. Compared to the cassette-based system, digital audio technology offers significant improvements to patrons of the program, network libraries, and NLS. NLS has therefore determined to implement digital audio technology as the framework of the future system. Following a decade and more of research and the establishment of solid technical foundations, the actual transition to patron use of digital materials is expected to begin in four years and require approximately five years to complete.

NLS anticipates that, ultimately, most patrons of the program will download digital reading materials via the Internet using a variety of broadband channels, and store and read books and magazines on portable playback machines designed specifically for use with such a delivery system. However, several major technological and economic impediments related to the provision of broadband Internet access to patrons of the program are expected to continue for some time. Therefore, NLS has decided that reliance on such a delivery system must be postponed until technological advances make it economically feasible.

The plan for the medium-term implementation of a digital system will, therefore, require the mailing of a physical object to patrons, which they will load and play in playback machines; in this regard this plan shares some features with the distribution system presently employed for cassettes. NLS has investigated three types of digital media: (1) CD-ROM; (2) magnetic hard drive; and (3) flash memory. Based upon an evaluation of the relevant technological and economic characteristics of these three media types taken in tandem with the operational environment of the program, NLS has concluded that at this time a flash memory-based delivery system is the best alternative, with one major proviso: the wholesale unit price of flash memory must decline further before implementation is economically feasible. NLS is confident that this economic threshold will be reached by FY 2008. While CD-ROM and magnetic hard-drive technologies are both more mature than flash memory, neither would support a book and machine system sufficiently reliable for the application; both books and machines would be likely to sustain damage in normal use and machines in transit though the mails because of their inherent physical characteristics (the machines have precision parts built to fine tolerances and are extremely sensitive to shock, and CDs are inherently delicate). These technological Achilles heels would result in higher media and machine failure rates than with flash memory, lead to poorer service to patrons and extra workloads for network libraries, and necessitate relatively high book-replacement, machine-replacement, and machine-repair costs to NLS.

NLS envisions numerous advantages, relative to the present system, from the implementation of a flash memory-based system. Audio reproduction quality is superior in flash memory, which will provide better sound quality in playback; flash memory has higher storage densities and capacities than cassettes, which will provide better portability for patrons; there will be fewer items for patrons to keep track of (generally one cartridge instead of several cassettes), and there will be no need for patrons to flip a switch or turn the cartridge over during play. Flash memory will require less storage space for collections at network libraries. The memory cards are long-lived and can be reused many times without loss of quality. Playback devices will have smaller physical size and weight and better portability than cassette players and require less storage space in network libraries. The machines will be more reliable than cassette machines, resulting in fewer malfunctions and fewer repairs by network agencies and NLS contractors; they will be easier and less expensive to maintain, requiring simpler repairs by network agencies, volunteers, and NLS contractors; repair parts storage requirements will be less, and parts expenditures reduced; and the playback devices overall are expected to be less expensive, enabling savings to be redirected to other facets of the program's operation.

The proposed structure of the future system will be based on the "Specifications for the Digital Talking Book," ANSI/NISO Z39.86-2002, which was approved March 6, 2002; digital books or machines, for any delivery system, must be able to process files based on this standard. Because development of the digital collection is independent of the medium to be employed in the delivery system, and with the digital talking-book standard in place, NLS has been developing its digital collection since FY 2002. NLS will produce all recorded book titles in digital format beginning in FY 2004, and has goals to produce 10,000 new titles in digital format by FY 2008, convert another 10,000 existing analog titles to digital format by FY 2008, and continue to produce an average of 2,000 titles per year during FY 2008 and in future years. The medium to be used for digital books is expected to be a high-density 128 MB flash memory cartridge, or card. A single card would be capable of holding the digital files for an average NLS book (approximately 12 hours), having been compressed using the AACPlus algorithm. The card would be physically larger than a credit card, but smaller than a cassette, and would be labeled in large print and braille. The present cost of such a card is approximately $25-$45 and will have to decline to approximately $10 to be economically feasible; however, NLS is confident that price reduction trends will continue and that this level will be reached by FY 2008.

NLS has determined that a Pareto's Law effect applies to book circulation in the free library program, i.e., the most popular 20 percent of titles constitute 80 percent of total circulation. Given this consideration-taken together with operational, technical, and economic characteristics of a flash memory book storage and distribution system and the traditional responsibilities of the parties who operate the program-NLS is proposing a "hybrid" business model for future operations to maximize the system's service effectiveness and economic efficiency. In this model, titles expected to be among the most popular 20 percent, hence constituting 80 percent of total circulation, would be mass-produced, allocated, and distributed to network libraries, and then stored in collections and loaned to patrons, much as cassettes are at present. The remaining 80 percent of titles, constituting 20 percent of total circulation, would be duplicated on demand by an NLS contractor operating a highly automated distribution center that would loan books to patrons nationwide. No book collection copies would be stored in the centers; rather, books would be stored on one or several mass-storage (multi-TB) servers located on-site and loaded on to cartridges at high speeds. Network libraries would provide reader advisor services for all titles, so there would be no contact between the storage centers and patrons; libraries would telecommunicate book orders as necessary to the center, which would confirm in the same manner both the issuance and return of books loaned to the libraries' patrons. The centers would reuse the cartridges many times; unneeded book copies in network libraries would be periodically weeded and allocated to other libraries in need of those titles, reused for mass-production of books, or reused for duplication-on-demand by the centers.

NLS expects to have 50,000 digital playback machines available for distribution by FY 2008. The plan is to have two types of machines, the predominant type intended for straight "linear" reading and the other with more-complex features. Many aspects of machine operations in the future system will remain unchanged, including distribution of machines from manufacturers to network libraries; allocation of production based upon readership; storage, lending, and control of machines by network libraries; management and procurement of repair parts by NLS; and the proportion of attrition associated with lost and stolen machines. Several aspects of machine operations will change, however, including lower disposal rates thanks to higher reliability (including better spill resistance); lower failure frequency and repair rates because of higher design reliability; lower repair costs because of better maintainability design; and relatively less reliance on volunteer repair and more on commercial repair.

Prior to distribution of the first digital talking books (DTBs), NLS will instruct network libraries to remove from their collections, if they desire, their entire rigid disk (RD) collections, to be disposed of or transferred to the NLS multistate centers, which will meet residual RD circulation requirements. Space now used for RD storage will be available for digital book storage and other uses. Cassette books and machines will continue to be the backbone of the system through approximately FY 2007, but in the five years between FY 2008 and FY 2012, digital books and machines will effectively replace cassette books and machines as the primary technology employed in the system. At some point during this conversion period, it will become too expensive to mass produce cassette books in small quantities, and production will cease. Copies of cassette books will be weeded from the library collections periodically as popularity wanes. The disposition of the cassette machine inventory will be dictated by the course followed by cassette book collections and demand. The modus operandi to be employed for digital magazine storage and distribution is yet to be determined; a disposable flash memory-based magazine will probably not be economically feasible. Two alternatives under consideration are returnable magazines and a delayed digital conversion of the magazine title list. It is likely that magazines will be produced on disposable cassettes well beyond the end of the book and machine conversion period.

In the business plan, the estimated major NLS-incurred costs of future book and machine operations for the hybrid business model are developed and compared with two other business models utilizing flash memory media and with the costs of current operations; these estimates and comparisons are illustrated in the business plan. The alternative business models considered for future operations are, first, one in which all digital book titles are mass-produced and distributed to network libraries, and, second, one in which centralized circulation is provided by NLS. The major functions of book and machine mass production are developed for both current and proposed operations, and the costs for the NLS book distribution center(s) are also developed for the proposed operations; the levels of demand for books and machines in the future system are assumed to be the same as for the current system. Assumed values for all factors influencing costs are provided. A total of $23,148,000 per year of NLS-incurred costs is estimated as the baseline average cost for audio books and machines for the current system; the hybrid model is estimated to require a $3,121,150 initial investment (amortized over 20 years) and $22,528,000 per year; the all-titles-mass-produced model requires no initial investment and $30,280,000 per year; and the all duplicated-on-demand model requires $15,606,000 initial investment (amortized over 20 years) and $38,524,000 per year. Hence the estimated costs for the hybrid model compare very favorably with the costs for current operations, while the all-titles-mass-produced model is approximately 30 percent higher, and the all duplicated-on-demand model is about 65 percent higher. It was determined that total NLS costs are very sensitive to the proportion of network circulation handled by the centers, the unit cost of distribution at the centers, and the unit price of flash memory, while they are not relatively sensitive to cartridge loss rate and book turnaround time, assuming that the known ranges for these variables are correct.

The estimated cost impact upon network libraries of the implementation of the hybrid model is illustrated in the business plan, based upon assumed changes in cost components and FY 2002 NLS Life Cycle Cost model baseline values. Estimated FY 2002 baseline network costs of approximately $61,328,000 would be reduced by approximately $9,173,000, or 15 percent, by the implementation of the hybrid model. More than half the total savings would be realized in facility and occupancy cost reductions because RD collections would be removed, recorded cassette (RC) collections would be weeded down, digital books and their containers would be physically smaller, only 20 percent of digital book titles would be stored in library collections, and random storage of collections would (possibly) be utilized. The balance of the savings are attributable to the assumption by NLS centers of 20 percent of the digital book distribution workload (that has been traditionally performed by network libraries) and to the higher reliability of digital books and machines relative to the current system. USPS-incurred costs are not considered in the analysis because the USPS is indifferent to the delivery system employed as long as materials transported though the mails are packaged, addressed, and identified in accordance with specified postal standards. The business plan presents a preliminary projection of long-term funding requests by NLS for implementation of the digital system. NLS intends to pay for modifications of the READS circulation system, but network libraries must pay for modifications of other information systems used that are necessary to support future operations. If the unit price of circulation from the centers is significantly higher than $1 per copy, the possibility of network libraries reimbursing the centers may have to be explored.

There is acknowledged pressure at both federal and state levels to facilitate more efficient operations through the implementation of new technologies. This pressure is sometimes answered by "quick solutions" with the primary objective of reducing short-term operating costs. NLS has been asked why the implementation of digital systems is not proceeding more rapidly. The answer to this question is twofold: first, the transition is too complex to rush and requires careful planning and execution; second, the unit price of media in the flash memory-based system must decline. Therefore, the fundamental transition strategy is to await further price reductions and, in the interim, execute all other necessary planning and implementation steps.

Decisions made in the medium term do not preclude the implementation of other alternatives in the future as the fluid digital environment evolves and new or significantly altered technological options become available.

The business plan contains a Gantt chart of the initial transition plan to effect implementation of the proposed flash memory-based hybrid model system, which shows the major transition steps and their approximate timing. Some of the first steps have already been taken. The major conversion steps will occur beginning in FY 2008 and continue through FY 2012, and the last steps will continue beyond FY 2012. The major steps in chronological order of their commencement are:

development of digital book title master collection-conversion of 10,000 existing titles and production of 10,000 new titles by FY 2008;
user survey-assess program patron characteristics;
player transition study--determine the appropriate long-term phase-out strategy for RCs/Cassette Book Machines (CBMs) and phase-in of DTBs/digital talking book machines (DTBMs);
pilot test of Internet delivery, beginning with magazines-learn lessons for the future system;
distribution media and system study-evaluate alternative media and select the best for implementation;
DTB player and media design contract-single contract to develop complete and coherent digital book and machine system for selected medium;
DTB container design contract-design of the digital book mailing and storage container;
distribution system design and transition planning-evaluate alternative distribution systems and develop detailed operational specifications and cost estimates for the selected system;
RD phase-out-remove RD collections from network libraries, dispose of some, transfer others to multistate centers;
circulation systems redesign-modify existing library circulation information systems to support future operations;
circulation systems implementation-install, test and implement modified information systems;
distribution system implementation-commence operations of the future distribution system;
DTBM production-mass produce digital playback machines;
reduction of CBM inventory-reduce and phase out cassette book machines;
medium production-mass produce cartridges specially modified for program;
DTB container production-mass produce digital book storage and mailing containers;
DTB mass production—mass produce digital books;
RC mass production phase-out-reduce and phase out mass production of cassette books;
weeding of RC collection-reduce and phase-out cassette book collections in network libraries;
weeding of DTB collection-weed digital books in library collections and transfer to other libraries or reuse for mass production or for duplication on demand.

A complete copy of the NLS Current Strategic Business Plan for the Implementation of Digital Systems is on the NLS website: www.loc.gov/nls/businesplan/index.html.

Library of Congress Home NLS Home Comments about NLS to nls@loc.gov

About this site Comments about this site to nlswebmaster@loc.gov

Posted on 2006-05-30