National Library News November 1999 Vol. 31, no. 11

Digitization at the National Library: Some Lessons and Outcomes

The National Library of Canada’s digitization activities began in 1995, when projects such as Canadian Confederation and Celebrating Women’s Achievements were developed and launched. Today, these two Web sites are among the National Library’s most popular and well-used digital resources. Building on these early successes, the National Library of Canada (NLC) has continued with the digitization of printed texts such as books, periodical articles, indexes, manuscripts, illustrations, photographs, maps and sound recordings, all drawn from its collections.

The NLC’s choice of digitization projects has been very selective and has emphasized adding value by improving users’ access to materials and by providing contextual information as a basis for interpretation and understanding of the digitized materials. Among the key criteria used in selecting materials for digitization has been a consideration of the importance of the material and its value to users. In digitizing these important Canadian materials and making them widely accessible to users via the Web, the National Library aims to contribute to the growing number of national digital resources now available.

Over the next three to five years, the NLC will continue with selective digitization to create a well-focused, accessible, retrospective digital collection. This will complement its collection of current electronic publications and give users a view of the richness and depth of NLC’s resources. Thematic, reference, and collections components will be created dealing with Canadian literature, music in Canada, and Canadian history and society. Appropriate in-depth resources, based on NLC’s unique and most significant collections, will be created where need and user interest are demonstrated. In connection with these will be the creation of resources useful in reference work and resources supporting NLC’s other services.

The NLC is proceeding on this path having reflected on some of the lessons learned from its digitization efforts to date. While it is impossible in a short article to do much more than quickly skim over a few of the key items learned, the following represents a distillation of some of NLC’s experiences:

• Audience:
  The audience and their needs must be given careful thought. This applies as much to the selection of the project and materials as to the design of the navigation within the site and to considerations such as file size, and the proper use of the ALT tag to ensure that the site is accessible to, among others, those who are visually impaired. The needs of the scholar, the general public, and students in elementary or high schools all differ significantly. These considerations extend to how the audience will gain access to the Web site and what constraints their equipment or Internet access may impose. For example, vibrant, detailed graphics can look very attractive and contribute much to the overall appeal of a site, but unless the user has a high-speed modem to provide access, the lengthy download times for large graphic files may prove frustrating.
• Copyright:
  Three to four months of lead time are needed in order to check copyright ownership and obtain any necessary permissions before the actual work of digitization begins. If materials cannot be used, either because permission is refused or because acceptable terms cannot be reached, this lead time provides an opportunity to select alternative materials for digitization. For some of its projects, the NLC, like many other institutions, has chosen to concentrate on the digitization of older items which are in the public domain. While this is a useful approach, inclusion of at least some contemporary materials is often necessary to provide an up-to-date context for interpreting the public domain materials.
• Selection:
  Given the volume of material which could potentially be digitized, the limitations of human and financial resources, the realities of copyright considerations, and the importance of complementing and avoiding duplication of other digital resources/digitization projects, selectivity is essential. The more care which can be given to the selection of materials or collections, the more useful the final result will be to the intended audiences.
• Check assumptions:
  To the extent possible, check all the assumptions before starting a project, or, failing that, early during the start-up phase of the project, when changes can be made at minimal cost. This applies to considerations such as the predicted cost of digitizing items, whether copyright permissions can be obtained and the effort required to do so, error rates, and the ability of staff or contractors to properly apply standards and best practices.
• Content:
  Make sure that the content experts are free to contribute their best by providing sufficient technical support for projects. This will allow them to focus on ensuring that the content is as clear, well-chosen, and well-presented as possible.
• Project management:
  Anticipate project overheads with care and plan for contingencies. Through what some may call a curious twist of fate, even small projects involving a few contractors appear to require considerable effort in their planning, administration and evaluation. An important part of project management involves checking the work which is being done to ensure that it conforms to the project specifications and to standards and best practices.
• Project team:
  Establish a well-balanced team for each project. Some of the assets that team members may bring to the project include project management skills, subject expertise, markup skills, programming skills, writing/editing capabilities, and graphics designing skills. Not everyone will have the same expertise or abilities, but team members’ skills and knowledge should be complementary.
• Maintenance:
  Document how the digitization has been done and the design choices made, and ensure that this information will stay with the site and be easily found and accessed in the future. Plan for the resources to maintain the project site after the work is officially over. This has implications for the project team. If temporary staff or outside contractors are used, at least a couple of the regular library staff must be capable of maintaining the technical and content aspects of the site to ensure that it remains accessible and up-to-date. This may mean temporarily replacing a permanent staff member so he/she is available to work full time on the digitization project team. This will ensure that, at the end of the project, there is the necessary blend of technical and content expertise needed to maintain the site.
• Planning:
  It is crucial that the amount of work involved in digitization not be underestimated. There are always last-minute changes, debugging, design changes, translation, final editing of content, checking links and technical quality, and proofreading to be done at the end of each project. Sufficient time needs to be built into each plan to account for these needs.
• Costs:
  The NLC has kept close track of digitization costs, which have varied considerably from project to project. These cost variations arise from a number of factors, including the following: the number of items being converted; the age of the materials and their physical condition; the number of different formats of items involved (e.g., books, manuscripts, periodicals, newspaper articles, photographs, lithographs, and 78-rpm sound recordings); the amount of specially formatted texts to be digitized (e.g., tables of text or numbers); the needs of the intended audiences; the subject and technical expertise required; the amount of staff training necessary; salary costs; linguistic considerations; and the need for systems development or special equipment and software. NLC costs range from approximately $2.00 per page to convert late 19th-century Royal Commission reports to PDF image format, to $6.50 per page for circa 1970 bilingual texts with a high proportion of half-tone monochrome illustrations, to $10.21 per page for fully encoded text (of about 500 words per page).
• Partners:
  Partners can bring additional skills, expertise, human and financial resources and unique materials to digitization projects. This results in projects which are more feasible and which have a "critical mass", making them more attractive to potential sponsors. In addition, partners can often also bring a different perspective to a project, resulting in fresh ideas, and results which are more interesting and useful to the intended audiences.
• Evaluation:
  Some form of evaluation and user feedback is needed for each project. Ranging from anecdotal or modest to extensive and sophisticated, the evaluation can be performed during the project, at its end, as an ongoing concern, or as part of an overall evaluation of several projects. It is clear, in any case, that no single measurement or methodology is sufficient to adequately evaluate a Web site. Instead, it is necessary to combine and balance approaches. For example, merely counting the number of "hits" drastically inflates the perceived usage of a site. Other measures of usage, such as Web page requests, can give a firmer sense of a site’s usefulness when combined appropriately with other approaches to evaluation.
• Metadata:
  The process of digitization affords excellent opportunities for the creation of metadata, especially if the person digitizing (converting, coding, proofreading) a given document also creates the metadata. There is more opportunity to become familiar with the content and, with proper training and experience, this person can readily apply the knowledge gained to create the metadata. Automation in support of metadata creation is essential to increasing the efficiency of the work and the accuracy and consistency of the records, whether they reside in a database or as embedded metadata within documents.

These are some of the lessons learned from the past five years of digitization. As the National Library of Canada moves into its next phase of developing digital resources, we would welcome any insights which the rest of the library community may have into these processes.