American MemoryThe National Digital Library Program:
Archived Documentation

The Library of Congress / Ameritech National Digital Library Competition (1996-1999)


Lessons Learned: Workflow and Project Management

Several awardees mentioned that they needed to adjust workflow and the assignment of tasks in light of experience. Quality control and housekeeping tasks, such as copying and backing up files, often took more time than expected. Some institutions mentioned the importance of establishing regular channels of communication among team members. The need to allow for contingencies was also stressed. Those using outside scanning vendors emphasized the need to work with the contractors to achieve the desired quality of scanning and understand the technical options. In-house activities could be delayed through problems with equipment or software.


University of Chicago

Online Collection: American Environmental Photographs, 1897-1931

This collection consists of 5,800 photographic images in various original formats and is being scanned in-house. Item-level descriptions are being prepared during the project period.

In the first interim report, Alice Schreyer emphasized:

  1. "One of the greatest challenges of this project is the need to ensure communication among staff across divisions, and the need to ensure that staff participating in the project at various levels are well informed and can contribute productively to the management of the project. Two complimentary means have been established to accomplish this goal: project management meetings scheduled on a periodic basis to update staff and to reach consensus on action items, and an e-mail reflector to facilitate discussion on project questions and issues."
Duke University

Online Collection: Historic American Sheet Music

3,000 pieces of sheet music have been scanned from originals in-house. In the final report, Steven Hensen noted that:

  1. ". . . the importance of workflow control, data file management, routine maintenance of the scanners and workstations, and the importance of automated processes and programming in completing projects of this size. Since all of the image data could not be stored on a single hard drive, the project manager organized the workflow in such a way to continually free the limited disk space by moving files to Apex disks and creating tape backups of their contents. The critical importance of frequent backups of these disks was demonstrated by the loss of some images due to disk corruption. Fortunately these losses were minor and did not set the project back appreciatively. Likewise, because of the large amount of file creation and deletion activity, it was found that regular scanning and defragmentation of all the hard disks involved was required."
Ohio Historical Society

Online Collection: The African-American Experience in Ohio, 1850-1920

In relation to workflow and assignment of tasks for their in-house scanning and application of subject headings, George Parkinson emphasized:

  1. ". . . that more time than was anticipated was needed to back-up and to move files around, to rename files and enter data into the database. In addition quality control took more time than anticipated because of poor microfilm and paper quality. Many items have not scanned as cleanly as we had hoped, particularly microfilm. Several newspapers were microfilmed with folds, were on poor quality paper originally or had print with broken fonts. Thus, it involves a significant amount of time to determine scanning levels which will allow for both text and images to be visible."

  2. "Originally, it was planned that the Scanner Operator would do the bulk of the scanning and that the Assistant Cataloger would focus on assigning subject headings to all the materials. However it was determined that assigning subject headings progressed more quickly than anticipated and that scanning was slower than originally thought. As news papers form the bulk of the material to be scanned, it was decided that the Assistant Cataloger will focus on scanning news and the Scanner Operator will focus on scanning and completing OCR on the serials, then manuscript materials. Both will work scanning microfilmed newspapers and assigning LCSH to them."


In the OHS final report, Parkinson discussed:

  1. "One of the most valuable lessons learned was the need to allot more time and resources for such projects. In previous digital projects, staff had worked extensively with one type of material. In every project, there is a learning curve and necessary time for adjustments. Project staff drew upon their previous experience when estimating rates of work. However, this project involved numerous different formats, [and each format required] many new decisions and long learning periods. During the course of the work, we realized that insufficient time had been allotted given the diversity of formats."
University of Texas, Austin

Online Collection: The South Texas Border, 1900-1920: Photographs from the Robert Runyon Collection

The 8,241 photographs in this collection have been scanned onsite by a contractor. In the first interim report, Mark McFarland noted that:

  1. Image quality problems were encountered due to technical problems inherent in the 8-bit digital camera used by the vendor. A quality control assistant was assigned to work directly with the vendor. A new 16-bit camera has greatly improved image quality and reduced the time needed for quality review.

In the final project report, Harold Billings discussed:

  1. "Even under what we considered to be optimum conditions for managing a digitization project, we encountered difficulties we did not anticipate. . . . The digitization process took twice as long as we anticipated due to staff turnover within the [ contractor] firm and due to the firm's mid-project acquisition of new scanning equipment that we elected to allow them to use on our content. This is not something we could control, yet it had a large impact on the project. Combined with, and in part, caused by the unexpected changes in personnel, we experienced several quality problems that took a great deal of time to isolate and correct. . . . [Also] the physical environment in which image capture was done was not wholly conducive to the production process. There were subtle camera movements caused by vibrations in the wooden floor in the scanning room and imperceptable electrical surges [that affected image quality]."
Graduate School of Design at Harvard University

Online Collection: American Landscape and Architectural Design, 1850-1920: A Study Collection from the Harvard School of Design

This collection consists of 2,500 lantern slides. A contractor took 35mm photographs of the slides and converted the images to Photo CD. Derivative images were created by library staff.

In the Final Cumulative Report, Ann Whiteside commented that:

  1. "The lantern slides were not un-mounted from their glass before shooting and scanning took place. We felt that un-mounting the slides would not give us any advantage in the scanning process, and would add to the project time and costs because each slide would have to be re-mounted afterward. I think that this decision was a good one, though when we looked at 35mm. slides and scans we received from [the vendor], we found that many had been shot too close to the edges of the images and images were cut off at the edges. We asked [the vendor] to re-shoot and re-scan approximately 150 of the 3,000 lantern slides. In addition, there are about three dozen hand-colored images in this collection which we had to have re-shot and re-scanned a s well, for color correction. In a future project, I would engage in more extensive discussions with several vendors before choosing one for scanning a project such as this. I think there are issues that need to be laid out very clearly, such as what film type to use for interpositives, exact parameters for framing images when they are made into 35mm. slides, and scanned, and any issues around color images and the original film.

  2. "We encountered some technical problems at the Graduate School of Design during image processing. We had a server that went down several times due to problems or regularly scheduled maintenance. Though we allowed two weeks for this process, it in fact took four weeks. I would say that in the future some of the issues we encountered this would not be issues (such as planning maintenance and image processing activities more carefully), and others would remain and should be accounted for in planning (this includes allowing time for unexpected problems)."

  3. "The research that went into the text portion of the document took longer than expected, as did the conversion to HTML format and the design of the web site material. In future projects, I would allot more . . . [time] for this."
Duke University

Online Collection: The Emergence of Advertising in America, 1950-1920

This collection consists of 8,500 images of various original formats and is being scanned in-house. In the interim report, Ellen Gartrell remarked that:

  1. "One of the lessons learned from this project thus far was the benefit of the selection of items prior to the beginning of the scanning process. Having time to ramp up to the project allowed for evaluation of selection resources and criteria as well as scanning/database procedures for each collection. Although there is always need for strategic readjustments throughout the project, a solid organized base from which to start is important."


  2. "Worklow control, data file management, and quality control are always important, and with a large student work force . . . these aspects become critical. Continual quality control, re-scanning instructions, and follow-up are necessary to maintain a consistent level of student assistant scanning output. Reviewing student scans as soon as they start work on each collection proved especially helpful, as it allowed the individual student to clearly see the standard expected before they incorrectly scanned a large number of images. Re-scans were still necessary, but the number was cut down significantly after the students were educated through trial and error at the start of the collection."
Brown University

Online Collection: African-American Sheet Music Digitizing Project

This collection consists of 1,500 pieces of sheet music that were scanned off-site by a contractor. In the third interim report, Rosemary Cullen observed that:

  1. While the staff had not previously undertaken a digitizing project, their considerable experience managing other types of grant-funded projects was enormously helpful, as were the contributions of colleagues from several divisions, including cataloging and systems. Among the factors that enabled the project to run smoothly were "The physical materials selected were simple and easily handled: flat sheet music that could readily be placed on a scanner without harm. Project staff could quickly be trained in safe handling. The materials were divided into color images and black and white images, facilitating the establishement of routine workflows for the vendor. The existence of full, high-quality bibliographic records for the originals meant that, with brief training, project staff could edit and clone the records for digitized images readily."
Michigan State University with Central Michigan University

Online Collection: Shaping the Values of Youth: A Nineteenth Century American Sunday School Book Collection

This collection contains digital images and encoded electronic texts of books. In the interim report, Michael Seadle noted that:

  1. "I would advise people to do more space planning at the start of grants -- both computer and human. Computers become significantly less efficient when their storage fills up, and people, even students, do better work in ergonomically optimal conditions." The impact of server-space needs for concurrent digitizing projects and semester-time demands for floor space in the Library building had not been anticipated adequately.

  2. "We chose to do the keying ourselves, rather than outsource this task, as many others have done. . . . Although this has worked out well for us and for our students, it has meant extra planning and extra local commitment. No one should expect to take on this sort of work casually."
Lee Library at Brigham Young University with the Utah Academic Library Consortium and the Utah State Historical Society

Online Collection: Pioneer Trails: Overland to Utah and the Pacific, 1847-1869

This collection consists of 6,040 images of various original formats and transcriptions of handwritten diaries. In the interim report, Susan Fales notes:

  1. One lesson learned "centers on the level of review of the original materials. We discovered that the criteria we had established for the selection of diaries and journals was not detailed enough. Several of the journals, which were identified through cataloging and finding aids as meeting the criteria of an original diary written on the trail during 1847-1869 and not previously published, did not qualify for this project. Most of these diaries turned out to be reminiscences or had been copied by someone else."

  2. "Most project planning centered on the scanning phase and not the delivery or even the packaging of the collection. The scanning has really turned out to be the easiest phase of the project, partly because we have people whose full-time job is devoted to the project. All of the other people involved are sandwiching the process among regular responsibilities.

  3. "The web design and packaging of this collection as a publication is where I think our vision was the shortest. . . . We are beginning to understand the issues which relate to bringing an anthology collection together in what is really a digital publication. We had paid absolutely no attention to this piece in the application process for the grant."