“3 out of every 4 visitors to the Met never make it through the front door” titled the New York Times on March 29, 2006 [1] in an article outlining the increasing significance of museum websites for the overall museum enterprise. Surveys by the Institute of Museum and Library Services (IMLS) on technology and digitization [2] reflect the growing importance of digital surrogacy in an upward tick of museums digitizing for access: while in 2002, only 6.1% of respondents identified better access to collections as their motivation, it emerged as the primary reason for digitization with a solid 56% in the 2005 survey.
As museums have learned to value the impact digitized collections make on their audience, they have also come to value the digital surrogate itself as an asset worth tracking and maintaining. This RLG DigiNews Special Issue “Managing Digital Assets in US Museums” shines a spotlight on first efforts to manage the growing number of digitally reformatted collections. Over the last 18 months, an overwhelming majority of the largest US art museums (represented in this issue by the Metropolitan Museum of Art and the National Gallery of Art) have started investigating digital asset management systems, with many smaller institutions (as represented by the Center for Creative Photography at the University of Arizona) joining the quest for the optimal way to handle the output of their digital photography labs and scanning operations.
While digital asset management systems come in many flavors, at their core, they provide the capability of ingesting, describing, tracking, and circulating a digital file. A number of themes recur throughout the articles of this special issue as the authors describe their journey towards implementation:
the disruptive nature of implementing a digital asset management system, which touches many stake-holder departments and demands a stringent policy framework on all levels
the motivation of supporting more efficient request fulfillment in-house as well as external access through the digital asset management system
the necessity of integrating the asset management system with various other key systems such as collections management or digital rights
the hope that the asset management system may provide a first step towards ensuring the long-term accessibility of the digital file
The last point may bear a little more scrutiny: the discussions around digital repository certification [3] have handily codified the belief that digital preservation is not primarily a function of technology, but a product of institutional commitment and policy. While systems will have to be capable of performing tasks such as file format migration, institutions will have to muster and declare the will to fund and sustain physical collections as well as their digital brethren. Digital asset management systems have raised the hope in the museum community that there may be an easy, off-the-shelf solution to the digital preservation conundrum, but their precise role in long-term retention remains uncharted.
In their early implementation, digital asset management systems have already proven to call the question on a variety of crucial policy and workflow issues. The three articles in this special issue trace the beginnings of these vital discussions in the museum community.
1. Vogel, Carol. “3 out of every 4 visitors to the Met never make it through the front door.” New York Times, March 29, 2006, Museums Special Section. Late Edition.
Museums large and small are investigating digital asset management systems (DAMs). Museum technology conferences offer workshops and sessions about digital asset management; our email lists buzz with chatter about how to select a system and questions about working with vendors and implementers. The Metropolitan Museum of Art—now at the start of the second year of a planned three-year digital asset management implementation project—was in the vanguard of American institutions choosing to invest heavily in digital asset management, and our team is consulted frequently about the Met’s project. Colleagues from other museums ask about our budget, about software and hardware selections, and about staffing and scheduling. We talk to them about the data model, and the relationship of asset information to collection information. Absent from the discussions we have had with these colleagues, however, are inquiries of a more fundamental nature—questions about the rationale, the motivation, and the policy framework that informed the Met’s decision to acquire and implement a digital asset management system. These issues, hotly debated by Metropolitan Museum staff and trustees in the years leading to the decision to acquire a digital asset management system, were, in fact, in the foreground of our institutional discussions. The thinking that resulted from the internal debates shaped the Met’s digital asset management project—now named Met Images—into an unusually effective collaborative effort, buttressed by a comprehensive understanding of the technical requirements for the project, but also by a broad institutional acceptance of the changes in policy and practice that will be brought by the work.
The Met Images project, in planning since early 2002 and under discussion long before that, is one of the most expensive non-construction projects ever undertaken by the Met. The three-year system implementation will, barring unexpected delays, be accomplished in a shorter period than the planning and discovery that led to the approval of the project by the museum’s executive staff and Board, reversing the more common ratio of planning-to-project years. The lengthy planning period was due, in part, to the Met’s commitment to asking and answering a series of questions not always built into the development process for digital asset management activities in the not-for-profit world. Our questions included quantitative ones such as: “How big is the collection of assets?” and “How much storage will be required?” We asked questions about proprietorship and access: “Who uses the assets?” and “Who will manage them?” We wondered about the quality and value of the assets: “Should images that are made available to the public be color corrected?” and “If the object’s descriptive record has not been reviewed, may it be distributed along with the asset?” and finally, “Are existing descriptions adequate to support successful searching at all?” We also asked a number of questions that forced the Met’s staff and executives to think through fundamental intellectual property policy positions: “Who decides who may use the assets, both inside and outside of the museum?” and “Is the Met’s goal to profit from the licensing of images, or to support an educational mandate for broad distribution?” All of these questions needed to be considered in light of a process that would, inevitably, seek to automate the answers. Processes and policies that had heretofore been entrusted to individuals within the organization would need to be formalized so that a system might manage them; decisions that had been made on an ad hoc basis now needed to be seen as patterns that formed policies. And as we found answers to our questions, the scope of the project inevitably grew.
Making the Argument: The Business Case
Research for the Met Images project began in late 2002 with a review of systems already in place in museums in the US and abroad. Knowing that a number of museums in the US, the UK, and elsewhere had completed or nearly completed the conversion of their photography studios to fully digital operations, and recognizing that several institutions had launched sophisticated e-commerce image licensing storefronts, we had assumed that at least a few organizations would boast end-to-end, fully integrated, automated systems for creating, storing, managing, cataloguing, and distributing images—perhaps integrated with collections management databases for object information. We were surprised to discover that though many institutions boasted a few of the elements of such a system—a robust front-end for licensing or a well-organized, fully-populated image database, for example—none of the organizations we talked to had achieved a complete end-to-end solution. Only a few other institutions seemed even to aspire to such an integrated system, perhaps because in most organizations no one department or individual bore responsibility for all of the functions of a system such as the one we envisioned. Nonetheless, at the end of about a half-year of research, the Met team had determined that such a tool set was, indeed, our goal. We knew from the start that the project would be expensive and thus would require the allocation of capital funding by the museum’s Board of Directors and Finance Committee. A request to the Board requires—at the Met as at most museums—a detailed financial request accompanied by a “budget justification.” The Met Images team considered several budget justification strategies in the course of our needs assessment planning and research. We began by crafting a budget justification based on a return-on-investment rationale: the system costs would be paid for through the revenue and efficiencies that the new tools would generate.
Early discussions with institutions that had invested heavily in digital asset management tools usually led to a conclusion that the initial investment in tools had been made on the basis of the definitive need for such tools to fulfill requests to license images to paying customers. Museums in Europe—with image licensing operations that functioned as part of the “commercial” arm of the museum—were, in some cases, equaling or bettering the Met’s licensing income with smaller image collections and smaller staffs. In addition, at around the same time as research into Met Images began, the Met’s image licensing operation was itself undergoing review, and the Image Library—the group responsible for circulation of the museum’s collection of transparencies, prints, and slides—had challenged itself to increase efficiency by responding to more requests, and by answering those requests more quickly. By adding to this challenge a strategy for licensing the museum’s images more aggressively to a more commercial audience, we calculated that we might increase both revenue and efficiency. The implications of this argument, however, were significant. A commercial image licensing agency is a demanding customer, requiring high-quality inventory, well-catalogued (and catalogued in a way that is meaningful for its intended uses) and available for searching online. To support the return-on-investment argument that we were—at that point—developing, we would need to strengthen the quality and quantity of available images and their cataloguing, in addition to providing the Image Library with tools to quickly receive, review, track, and fulfill requests. Planning for an return-on-investment business case forced the project team to create a project plan that included inventory development and cataloguing as parallel (and equally expensive) parts of the digital asset management strategy.
The Met’s Board responded to the initial project proposal by asking two questions that significantly reshaped the project plan and justification. First, they asked the team to consider whether the project might be staged in phases, allowing for a review of the requirements and goals at the end of each year. Second, they wondered whether return on investment or profit were in fact the right drivers for this project. They suggested that the project’s goal might, instead, be about essential investments in the museum’s assets and infrastructure that would support core mission objectives: to research, document, and educate. With this guidance in place, the project was re-structured into three phases: the first primarily about the acquisition and deployment of the repository; the second about the broadening of access to staff throughout the museum, integration with other museum applications, and the implementation of more sophisticated security structures; and the third about enabling access to assets through a self-serve, Web-based front end. Despite the shift from a return-on-investment to a mission-driven rationale, the project retained its inventory and cataloguing elements, building into each phase activities intended to develop additional inventory and enhance cataloguing. The decision to maintain this emphasis on content development has helped the project to achieve widespread support throughout the institution.
New Directions and Unexpected Consequences
Recasting the project as one about the museum’s mission forced the project team to consider some new directions for the work. The project, for which funding would now be approved in phases, began with a first stage that was broadly labeled “Preservation.” The team developed an argument for the importance of creating a system for the protection of museum assets, for the development of a shared standard for cataloguing of assets that would enable successful retrieval, and for a centralized database that would streamline management of the assets. We also discussed the digital preservation procedures and strategies that might be enabled to support ongoing management and “permanent” storage of our assets, but agreed that the landscape of digital preservation—both costs and procedures—was, as yet, too hazy to allow us to commit to a strategy. We felt comfortable with the position that, in a later phase of the work, we would revisit digital preservation questions, and that the implementation of the system would buy us time to monitor developments in the preservation domain.
The project’s new direction also allowed the group to ask some useful questions about the costs and value of enabling broad access to museum assets. A system such as the one we proposed to build would contain tools that would support access to and distribution of museum assets at a far lower cost than had been possible in the previous system—which combined a high personnel cost (human beings had to review, research, and respond to each individual request for access to assets) and a high material cost (analog images, in particular, had a limited “shelf life” and were easily lost, damaged, or worn out). When it became clear that the new tools would significantly reduce “transaction” costs associated with fulfilling individual requests for images, we asked ourselves whether it was appropriate to continue to offer the same fees that we had previously—charges that had been largely based on recouping the museum’s costs for fulfillment. The group felt strongly that for scholars, a class of users that the museum seeks actively to support, any savings realized by the implementation of a system that would reduce costs should be passed on.
This discussion led eventually to the development of the museum’s “scholars’ license” program, a fee-free licensing initiative (pdf) announced in 2005. Under the program, a limited number of high-resolution museum images (those already prepared for distribution) will be distributed via third-party image distributors for a range of scholarly uses, including print publication in low-run quantities, free of charge. The program seeks to streamline requests and distribution activities, reducing the administrative burden of fulfillment on both the museum and the requestor.
Finally, discussions initiated within the project team about the requirements for broadening access to the assets led to the development of a major new initiative not originally foreseen by any of the Met Images team members. Believing that a number of users—including those requesting images for licensing, but also Web visitors, internal staff, and others—searched the collection using descriptive terms not contained in our existing cataloguing, the project team began a discussion with staff in a number of departments, including the libraries, education, and curatorial—about meaningful ways to collect subject terms, or keywords for retrieval. The group examined options ranging from engagement of permanent staff for keywording, to data mining of existing content. Every solution appeared inadequate; some were also expensive. Inspired by the work done at the Fine Arts Museums of San Francisco to deploy volunteers with index cards to describe works of art in “lay” terms, the group eventually turned to a plan that anticipated the blossoming of interest in “social tagging”—the employment of Web users to provide keywords (or tags), to label content. The Met’s internal discussions led to a broader discussion within the museum community at large and ultimately to the formation of the steve collaboration, a consortial development and research project that is investigating the usefulness of social tagging to describe museum collections and developing open-source tools for collecting and processing tags.
In 2005, nearly three years after we first began our research into digital asset management, the museum’s board approved the project plan, funding the first one-year, phase of the project, for the acquisition and implementation of Interwoven’s MediaBin digital asset management software. The first phase of the project also included funding for scanning and cataloguing the museum’s archive of large-format transparencies as well as preparing a decade’s worth of digital images for loading to the new digital asset management system. The project, largely on schedule and under budget, has continued into an approved second phase and year. In the months that have followed the approval of the project plan, many of our expectations about the challenges of successfully initiating a digital asset management project have proved correct; many additional issues and questions, large and small, have surfaced in the day-to-day work of the project. A few of these matters are addressed in the second half of this paper.
What is the Content? Taking Inventory
The content that can be managed by digital asset management systems is voluminous and includes, but is not limited to, image files, sound files, video files, desktop publishing files, pdfs, and presentation slides. Because of the significant resources required to make assets available within a digital asset management system, institutions must evaluate assets and asset collections to determine whether (or when) they will be included in the system. Evaluation of these assets should include careful consideration of a number of attributes, including uniqueness, image quality, cataloguing quality, replacement costs, and centrality to the institution’s mission. Ideally, inventories should probably be taken on a recurring basis and concurrent with efforts to load identified collections to the digital asset management system. Though the inventory process can be labor intensive, it is an essential stage in minimizing redundant activity, identifying at-risk collections, and setting priorities for digitization, cataloguing, and ingest into the centralized digital asset management system. At the Met, we have created an inventory questionnaire to elicit information about “local” collections—the uniqueness of the materials, the size of the collection, the characteristics and format of the files, the extent of cataloguing, and the frequency with which staff members access the collection items. This inventory work has already served to identify departments managing valuable content with institutional or long-term value and also departments unaware that their material is derivative content duplicated elsewhere in a higher-resolution or better-catalogued version.
Having identified images of works in the collection as its highest priority assets, the Met resolved to load into the digital asset management system the entire digital production of the Photo Studio (more than 200,000 digital files, captured in a decade of digital photography and scanning) in the first phase of the Met Images project. We would also load electronic records of analog images and negatives stored in both the Photo Studio and Image Library. Collating these images and their data in a useful format proved to be a challenge. High resolution images had been stored on optical media with corresponding metadata in Excel spreadsheets. A first task was to move these images and their metadata to a medium more suited for networked communication. A disc publishing robot was reprogrammed from its intended use as a high volume DVD copier to a more benign use: grabbing and dropping discs into a computer to be copied to a network attached storage device. Once the files and data were copied, we began the process of normalizing the Excel sheets and evaluating the usefulness of the files. A significant amount of work was done to massage the data and files to ready them for ingest into the digital asset management system. And, although the assets have now been ingested, a great deal of cataloguing work remains before the assets will be catalogued to a level that will support reliable access and use by the public.
Who Owns and Manages a Digital Asset Management System?
A digital asset management system is a technology tool, requiring support or management by technical staff. It is also a content repository, and responsibility for content must be given to content creators and managers. And a digital asset management system is a catalogue, providing tools for cataloguing work that is performed by the institution’s cataloguing professionals. Finally, as an ongoing institutional investment with storage or management requirements that may be restricted by budget or human resources, decision making and prioritization of what is managed, catalogued, and stored in a digital asset management system needs to be made by those with responsibility for organizational policy and resource allocation.
These varied and sometimes competing management responsibilities did not readily fit into the existing structures of the Metropolitan Museum. For the project to succeed, management of both the images and the image repository needed to be distributed among staff from different areas of the museum, and new working relationships had to be developed between areas of the organization that had not collaborated closely before. The museum’s senior administration did consider whether any single department might head this effort but decided instead to create a project team, headed by the Director’s Office and Information Systems and Technology and advised by members of the Image Library, Photo Studio, Counsel’s Office, and Finance Department—collectively, the Met Images “Lead Team”—to manage the Met Images initiative while in its implementation phase.
The Met has hired a project manager in a three-year temporary appointment to facilitate the many aspects of the project. The Manager of Met Images is assisted by staff responsible for technology, content, and the cataloguing and circulation of images. Together, these staff members are defining new tasks and workflows for content creation and management, as well as forging cross-departmental relationships for developing, utilizing, and managing the system. Defining the new tasks and workflows in each of these areas has also illuminated institutional policy questions requiring decision making taken in concert with the project’s Lead Team. When the system and its policies and procedures are in place, the new tasks and workflows will be ceded to their appropriate permanent administrative areas, specifically: technology to Information Systems and Technology; inventory and content creation to Photo Studio; and cataloguing and circulation to Image Library. What follows is a more detailed description of a few of the tasks, workflows, and issues considered by the three administrative areas and the Met Images Lead Team.
Some Key Issues and Concerns
Technology
A well planned and managed IT strategy is central to digital asset management. Managing a digital asset management system requires contributions by IT staff with a range of different skill sets, including networking, database administration, Web server administration, storage and backup strategy, and applications support. The Met Images Lead Team believed that an experienced information technology generalist to coordinate the efforts of the various IT staff working on the project implementation would be needed; as soon as funding was secured, a dedicated analyst was hired to coordinate the day-to-day technical aspects of the project.
For most institutions, implementing a digital asset management system signals a decision to begin storing larger files on the network in order to improve access to files that have traditionally been stored on less accessible and less secure media such as optical discs or external hard drives. The storage implications of a digital asset management solution can be significant. At the Met, we moved from a storage capacity of about seven terabytes of data on our entire network, to a capacity of 60 terabytes—expected to be adequate for about two years. While our storage needs are greater than those of most art museums, even at a small- to medium-sized organization the prospect of needing to increase storage on this scale in a short time can be sobering.
Image Creation
The museum’s content creators perform the essential function of populating the digital asset management system with assets. Imaging staff in the Met’s Photo Studio are responsible for creating images and contributing them to the digital asset management system. The key members of the imaging team at the Metropolitan Museum are digital photographers, scanning operators, and post-production staff. In their creative role, this staff is responsible for producing images of a very high standard. In addition, they support the administrative functions of the digital asset management system by producing a minimal level of cataloguing, capturing metadata specific to the image creation and editing process and also validating or creating the data that also allows the digital asset management system to match images to their corresponding object record from the collections management system.
Long before the Met Images project was launched, a well-defined strategy for conversion to an all-digital photography studio was underway. However, the studio was forced to engage with a series of questions not previously considered as preparation for the Met Images implementation moved forward. As the museum’s staff at large embraced the concept of a searchable, centralized repository of images, they began to wonder about the relationship of the repository to new photography ordered in support of museum programs such as exhibitions and publications. Often produced on a rush basis, such programs would be well served by up-to-date information about the status of photography orders and, better served still, by the ability to access and download new material as soon as it was ready. The existing photography workflow needed to be overhauled in order to integrate it fully with the digital asset management system and tools. This project, related but separate, constituted a major initiative in itself, and because many of the same staff members who were already working on Met Images would be involved, the work has been scheduled to occur during the second phase of Met Images.
In addition, the Photo Studio’s digital imaging standards—already well documented and tested—required revisiting since questions about the nature of an image file prepared for broad distribution were identified. The Studio’s post-production staff worked with the Met Images Lead Team to develop processes for cropping, color-correcting, and finishing image files and to institute standard descriptions for the quality and post-production state of the images.
Cataloguing
While creation and storage of high resolution digital images is at the center of discussion of many digitization projects, an issue of equal scale and complexity is the development and maintenance of high quality metadata cataloguing. The relative breadth, depth, and accuracy of a metadata record determine whether an institution will be able to answer the questions “What is the content?” and “Who may access it?” With the promise of collections of digital assets made available online through the Web comes the responsibility of developing and maintaining the documentation to make those collections most useable. A carefully constructed metadata schema can determine the true value of a digital asset management system to an organization’s internal staff and also to its external clients, users, and aggregators.
Metadata about digital assets related to collections of museums may be thought of in three categories: information related to the physical museum object depicted in an image, metadata related to the image itself, and data related to the rights and restrictions. In each of these areas, museums should look to existing and developing standards. Creating our metadata schemas based on these standards and creating crosswalks to them will allow us to better communicate with one another. Metadata is also an area where museums should proceed cautiously. DAM vendors and implementers are relatively new to the cultural heritage space. Cataloguing information about assets related to works of art is different from cataloguing information about sneakers or pickup trucks. The permanent and archival nature of our digital asset collections may require different functionality and data structures than those offered by most large commercial digital asset management vendors. The full relational content of our catalogue records from our collections management systems is not easily searchable in the flat file context of many leading systems. Nonetheless, as DAM vendors come to understand the requirements of our community better, we believe that tools will be developed that are more appropriate to the needs of museums, libraries, and archives.
At the Met, the requirements for the systematic management of metadata will result in a significant shift or growth in the cataloguing work of the museum’s Image Library, the department that has provided access to images of our works of art for nearly one hundred years. Knowledge about the rights and restrictions related to an image—previously implied by characteristics such as file location in the library or gift history of the work it depicts—now has to be made explicit as data. The Met’s Image Library will now assume responsibility for the task of ongoing metadata population, review, and augmentation. While the introduction of the digital asset management system will streamline some processes for discovery and circulation of images by Image Library staff, it will add a set of new cataloguing responsibilities to the work of the department, with attendant impact on staffing and training. Concurrent with the launch of the first phase of Met Images, the Image Library decided to hire its first Systems Librarian, a staff member responsible for training staff in using the new systems, documenting workflows, and participating in discussions about system and interface development.
Conclusion
The many and varied tasks that make up digital asset management represent a significant change to the way that the Metropolitan Museum goes about creating and sharing images of its collection. Although the management tasks are complex and sometimes unfamiliar, successful creation and adoption of them—paired with new technologies and platforms for communication of information about our collections—will expand the reach of our institution in a profound way. This much is probably true for most of the museums and cultural heritage institutions now considering the implementation of a digital asset management system. However, for each of these institutions, decisions about how to implement a digital asset management system, who will do the work, when to begin, and what assets to manage will certainly be based on informed choices at the institutional level. The decisions will, in the end, shape each organization’s digital asset management strategy into something that is distinctly its own, based on a process of institutional self-discovery that is likely to give birth to new alliances and collaborations, build new skills, and produce some outcomes that may reshape institutional policy and procedures in unexpected and fundamental ways.
“Digital asset management” is an often-used term in the information technology and media spheres. In the context of this article, a digital asset management system is defined as an application that gives both users and systems administrators organizational control and insight over various media file types and their associated metadata. Without a properly employed digital asset management system, file creation and storage can quickly become a jumbled maze of file system hierarchies, and search and retrieve strategies can be quick to fail.
For museums, there is clear business value in:
significant staff time-savings for file retrieval
alignment of appropriate files to particular end uses
elimination of redundant storage systems
management of image cataloging and storage
creation of an infrastructure for digital preservation
documentation of image provenance
linkage to other core systems that manage digital rights, collection information, or publishing
Background: Interlocking Systems
The Division of Imaging and Visual Services at the National Gallery of Art (US) provides institutional infrastructure and technical leadership in digital image capture, quality assurance, storage, retrieval, and delivery.
The Gallery’s transition from analog to digital imaging systems in 2004 necessitated the staged development of a comprehensive digital asset management strategy. To implement the first stage—management of high-quality direct digital capture of art—the Gallery acquired and configured Extensis Portfolio SQL. Currently, nearly all of the digital assets created or otherwise acquired by the Division of Imaging and Visual Services are managed in Portfolio, representing approximately 3.5 terabytes with a growth rate of about 100 gigabytes a month. Portfolio manages all file formats, versioning, and derivatives, as well as technical, administrative, and descriptive image metadata. It also supports many functional services at the Gallery, including review and management of extant color transparencies, seamless management of images for the National Gallery’s intranet and public websites, and interim work-order oversight.
The Gallery is now converting from a 6-part, dysfunctional paper-based work-order system into a browser-based Digital Image Request System. This new Digital Image Request System will support efficient image order management and request fulfillment for exhibitions, curatorial research, publications, education, art conservation, and a multitude of other museum programs and offices.
In addition to Portfolio and the projected Digital Image Request System, the National Gallery maintains a homegrown legacy enterprise collection management system running on an IBM mainframe. In late 2005, Gallery Systems was awarded a contract to upgrade and configure their product, The Museum System (TMS), to provide much of the functionality of the mainframe database. The functional gaps between the Gallery’s CMS and the current version of TMS will be principally bridged by a new enhanced version of TMS. Eventually all three systems, Portfolio, the Digital Image Request System, and TMS, will exchange data multi-directionally. The functionalities provided by these interlocking systems should be considered as core elements in a comprehensive suite of digital asset management tools and applications.
Image Repository
By developing the image repository in stages we were able to secure a low-risk, low-cost solution for an immediate need at the division-level for a digital asset management solution while deferring a long-term enterprise digital asset management system until we had other necessary pieces in place.
Portfolio met our requirements for a functional, off-the-shelf, cross-platform, multi-user digital asset management application. It is relatively inexpensive to deploy and easy to use and administer. And Portfolio is extensible—in our case with Applescript on the client side and PERL on the SQL backend.
Through the Portfolio SQL backend, integration with other information systems, such as TMS and the projected Digital Image Request system, is possible. If we transition to a major enterprise digital asset management application in the future, we should be able to easily convert our data.
Portfolio has been used to create four discrete image databases. These databases (called “catalogs” by Portfolio) are a) images of National Gallery art objects or extended loans, b) temporary exhibition objects, c) event photography, and d) technical research. The collection catalog also includes legacy digital images (images not color-managed prior to 2004) and images acquired from external sources for exhibitions and publications.
Portfolio manages all metadata in a SQL database backend and the digital image files are stored on a recently upgraded Apple Xserve RAID system (now 23.5 terabyte). Offline image collections that are currently stored on optical media will be transferred to the Xserve for easier access.
Apple’s file management technology has made implementing an accessible but secure repository fairly painless. File access via password is generally available throughout the Division of Imaging and Visual Services, with restrictions placed on certain collections. Write access (highly restricted) is granted on an as-needed basis and is primarily given to asset creators. In addition, Portfolio has its own password-protected security services that add another layer of read or write protection via access modes, known as administrator, publisher, editor, and reader. The administrator mode offers full access to all features. An administrator can assign passwords to other users, create custom fields, and control cataloging and startup options that other users can’t. Publishers can add, remove, export, and edit files. An editor can edit metadata. A reader can only search and copy.
Each master file is saved in the RAW image format at 16 bits per channel. All subsequent image editing is done on a full resolution derivative file in TIFF format with edits preserved in layers. Layers keep a saved history of any file changes, so files can be rolled back to their original states whenever necessary. The TIFF file is used to generate a third derivative file, our so-called “parent file,” with flattened layers. The parent file is used to generate other versions for specific output requirements. All three files (master, layered parent, and flattened parent) are saved and managed by Portfolio.
The use of industry-standard color management systems ensures image quality control consistent across all versions. This production environment enables a closed-loop color work flow and creates the “use-neutral,” high-resolution master files and a variety of derivative files that are created for specific needs (e.g., ink-jet and laser printing, pre-press imaging solutions, and Web publishing).
We maintain four primary Portfolio image catalogs corresponding to the four basic image categories referenced previously. In addition to the four primary Portfolio catalogs mentioned earlier, we have created several specialized Portfolio catalogs to manage specific projects or image collections while they are “works in progress.” To ensure consistency across catalogs, we have developed a standardized procedure for cataloging images and applying metadata to image records.
Image Ingest
The Master image is transferred to the Apple Xserve RAID server and added to the Portfolio image catalog by the asset creator, typically a photographer or a scan technician. The asset creator is responsible for assigning metadata as part of the ingest process. Image derivatives are automatically processed by Portfolio. The asset creator transfers them to the appropriate folder on the Xserve and Portfolio monitors these folders for changes and adds new images to the catalog.
For all images ingested, an essential metadata set (capture device, asset creator’s name, color profile information, date created, and EXIF metadata captured automatically by Portfolio) is appended to the image record. Supplemental metadata is also added to image records. The supplemental metadata is based upon the image category (e.g., permanent collection image, exhibition object image, non art image) and image type (direct digital capture or scan). Examples of supplemental metadata include: conservation treatment status [1], object collection information (from our collections management system), curatorial approvals, work order information, exhibition information, and so forth.
The Division of Imaging and Visual Services has also developed automated methods to validate image metadata and populate common Portfolio fields using information in the Gallery’s collections management system. The images are also automatically displayed on the National Gallery intranet and public Web sites via flag triggers. These can be refreshed or replaced with ease.
Additionally, there are plans to create custom XMP panels to trigger bi-directional data exchange between metadata fields in Portfolio and data in the file header, for example to embed a copyright notice, a URL to the National Gallery of Art website, photographer name, caption, and a unique object identifier number data from Portfolio into the file header.
Integrating Systems
The majority of digital images produced by the Division of Imaging and Visual Services are photographs of art, either owned by the National Gallery or on loan for exhibitions. Accordingly, integrating descriptive metadata from the Gallery’s collections management system with Portfolio was a key system requirement. Without descriptions, querying and reporting inside the asset management system would be impossible.
We developed a Web application that enables a user to easily retrieve object information from the collections management system and populate the image record in Portfolio. An Applescript within Portfolio activates the Web application, passing the image’s unique record ID number to the application and opening the appropriate URL in a Web browser. The Web page shows basic image record information retrieved from Portfolio, including the image thumbnail. To retrieve the descriptive metadata, the user initiates a search using the object’s accession number or unique object identification number. Object collection data [2] is retrieved from a regularly updated Postgres database extract from the collections management system. After retrieving the object information, the user clicks a button to automatically update the image record in Portfolio. Using the image’s record ID number, the application updates the record in the image catalog’s database directly.
Figure 1. Integrating systems.
Object catalog data is automatically refreshed each night in the Portfolio image catalogs. This refresh is activated via a stored procedure and requires no user interaction.
Intranet and Web Image Publishing
The Division of Imaging and Visual Services is responsible for publishing permanent collection and exhibition images on the National Gallery’s internal and external websites. The National Gallery publishes only “reference” view images. These are visible spectrum (i.e., not X-ray, UV, IR, or special technique) photographs of artworks currently not in any conservation state (e.g., during cleaning). Both conservation and reference images are managed within the same Portfolio catalog.
The first step in publishing is to submit content to the Gallery’s intranet. Visibility on the National Gallery’s intranet is controlled by metadata. Portfolio triggers image publishing to the intranet, a process augmented by a Postgres database extract of content from the legacy collections management system. A script within Postgres down samples a series of different-sized JPEG derivatives that are then displayed via searching or browsing. These files may be individually copied by all Gallery staff who follow usage guidelines.
A flag field in the collections management system controls visibility on the public website based on certain image use rights and permissions. If the image rights are held by or licensed to the National Gallery, the public website manager triggers external publishing. The following information is required to publish an image:
A unique object identifier: A key field, either accession number, object/child id number, or exhibition (loan object) id number that associates the image with the object record.
Image view: An image may either be a preferred or an alternate view. The preferred view is the image’s primary view – for paintings and other two-dimensional art this is usually the front full view of the object. An image may have only one preferred view; any other views of the object are designated as alternates. Alternate views must be numbered sequentially – this indicates the image’s position on the Web pages.
During development, the objective for intranet publishing was to control the entire process through Portfolio, which has proved to be straightforward and user-friendly. Several of the fields required for publishing (object catalog date, reference image, preferred view) are set during the ingest of the master. To publish, the user simply sets the remaining relevant field values and sets the flag field (i.e., “Publish”) to “Y” or Yes. The publishing process, like all imaging workflows, is very much a work-in-progress that will be further refined as new systems come online.
Figure 2. Publishing to the Web.
Current Enhancements
As mentioned earlier, the National Gallery is in the process of replacing its legacy collections management system with Gallery System’s TMS. Because TMS lacks a user-friendly image loading mechanism, image ingest and data validation is cumbersome. Therefore, we are developing a mechanism to link images to TMS directly from Portfolio. This process will work like our intranet publishing process. We expect to combine the existing publishing process with the new TMS image linking process into a single procedure that will allow staff to publish images to the intranet and load images into TMS in a single step.
Digital Image Request System
The objective of the Digital Image Request System is to replace the legacy paper-based, internal work-order system with a cross-platform (Windows/Mac) browser-based system. Objectives include:
Increasing production capacity with current staffing levels.
Increasing the efficiency of work order management and fulfillment.
Enabling the efficient exchange of data between the image request system and other financial, intellectual property, and collection management systems
Providing browser-enabled user interface
Providing direct and documented communication between requester and producer
Meeting expanded Division of Imaging and Visual Services program and management accounting and reporting requirements including:
Performance metrics
Order analysis
System status reporting
System activity audit trails
Service and product quality assurance
The processes of determining requirements, interviewing stakeholders, selecting vendors and applications, as well as integrating and implementing the system have required a tremendous effort by many departments throughout the National Gallery. Significant investments of time, interest, and technical expertise have been required of Division of Imaging and Visual Services staff and many others whose professional work requires the use of digital images, including curators, conservators, editors, Web and print designers, educators, registrars, press relations officers, librarians, and archivists.
The dividend of this effort was a carefully vetted set of system requirements, as well as a vision prototype—a graphical representation of the system functions. This prototype was useful as a shared document for stakeholders as well as a guidepost for contractors bidding on the project.
To accomplish the Digital Image Request System, the National Gallery acquired and installed Metastorm’s Business Process Management (BPM) software, a general purpose software development toolkit that is built around a graphics engine to rapidly develop forms-based applications and generate SQL database code. The software permits the applications developer, systems analyst, or system administrator to change rules, forms, roles, and workflows on the fly. Metastorm’s software has the potential to be used in a variety of National Gallery applications including work orders for rights and reproductions contracts, conservation, matting and framing, telecommunications infrastructure requests, and event scheduling. It includes templates that can be used to automate business functions performed by the human resources department, the travel office, and most other service areas in the National Gallery.
Next Steps-Integrating Rights and Reproductions
The Division of Imaging and Visual Services also manages rights and reproductions for the Gallery as part of its Visual Services operation. The functions of rights contract management and image licensing are fulfilled by a legacy system that runs on an unsupported application and operating system. A static page on the Gallery’s website describes the process for ordering image requests and licenses.
A critical next step after the implementation of the Digital Image Request System is to automate the rights and reproductions services. It is conceivable that a “shopping cart” model built into the Digital Image Request System could be adapted to the public side of rights and reproductions, with add-ons for invoicing, contract management, collecting receivables, and image delivery via FTP. Much of the Digital Image Request System work order fulfillment infrastructure is adaptable to rights and reproductions.
Finally, a strategic need that has been recently identified and broadly studied is to provide Gallery-wide intellectual property management for image uses (rights granted to or owned by the National Gallery, rights granted to others, credit lines, permission expiration dates, copyright owner information, use expiration, etc.). Contract histories with intellectual property owners (artists, estates, agencies, museums, stock houses) for image use for publications, websites, exhibitions, and education could be managed, described, and integrated in the image data management process. By managing this complex web of intellectual property functions, the National Gallery will be better positioned to respond to the ongoing and often tedious or redundant questions concerning rights to use or publish that often impede publishing workflows.
The Road to Enterprise Digital Asset Management
The current DAM was designed to provide the essential first stage of managing collection image assets at the division level. Enterprise Digital Asset Management (EDAM) can be thought of as Portfolio on steroids and available for all staff: an industrial strength application that provides direct access to images and image metadata throughout the entire museum. The EDAM becomes the umbrella under which to manage all the growing image resources used and maintained by other divisions (e.g., Conservation, Slide Library, Publishing, Exhibitions, or Archives). Additionally, the EDAM can provide a home for the multitude of research image files that are individually parked on desktops and servers throughout the Gallery without the prerequisites of descriptive cataloging and broad accessibility.
Many of these systems allow users to create and save sets of images through simple or compound queries and/or additive browsing. Other useful functions include the ability to save user-generated text and share it across the organization. This functionality can reduce redundancy and facilitate project management for exhibitions, research, and publications in every department. Investments can be further leveraged if bridges are built between these functions and TMS. Ideally, it shouldn’t matter where the image set building is initiated, either in TMS or the enterprise digital asset management system, but the end result is user access to critical sets of images that can be downloaded, edited, projected, or printed in groups.
Another important function that can be tied to EDAMs is image presentation. Preparing for a lecture using a mix of images from a variety of image sources, analog or digital, networked resource or private collection, is a daunting task. We need to provide a visual presentation tool for our lecturers that does not make extraordinary demands upon their technical ingenuity.
Meaningful discussions leading to the initial definition of requirements and scope of the National Gallery’s enterprise digital asset management system will begin in late 2007. It is clear that this project will grow and be redefined as dictated by institutional priorities and resources.
An enterprise system requires management and safe storage, access and delivery of all institutional documents, images, and other file types (audio, video, CAD). This could include the future management of department-level image management applications and functions that are either currently automated, in the planning phase, or exist to fulfill an immediate need. They include:
Archival records (e.g., historical documents and records managed by the Gallery Archivist)
Conservation documentation including treatment reports and scientific studies
Licensed images that are managed by the Department of Image Collections and used for internal scholarly purposes (e.g., Saskia and ArtStor)
Public lectures in audio and video formats
Staging image collections for distribution via FTP by demand or through an Open Archives Initiative model
More powerful automation of temporary exhibition related images for publication, education, and publicity
Cataloged storage in a preservation and access format for images that are created by non-imaging staff (curatorial assistants, conservators, registrars, plant managers, educators, etc.) with digital cameras [3]
Figure 3. Vision Prototype.
A full discussion of the potential of enterprise digital asset management systems and the practical intersections with the Open Archives Initiative is obviously beyond the scope here. The full architecture and implementation of enterprise-wide asset management is by nature a long-term continuum and will be a persistent challenge for museums.
Notes
1. Three conservation image metadata fields are employed with controlled vocabularies: Treatment (before treatment, during treatment, after varnish removal, etc.); LightSpectrum (visible, IRR, UV, multi-spectral, etc.); LightQuality (balanced standard reflector, overhead diffuse, raking, ratio 1:3, etc.).
2. The following collection management data fields are imported into Portfolio: Accession number, ObjectID, Exhibition number, Artist, Title, DateCreated, Dimensions, Classification, Medium, Credit, Location.
3. While this activity may appear to be “out-of-control,” it is essential that decentralized image capture be embraced. No institution can marshal the resources to fulfill all image documentation needs. It is important to provide a framework of basic practices for image creation, cataloging, and storage. This has been addressed at the National Gallery through workshops, Division of Imaging and Visual Services intranet resources, and active participation on the Universal Digital Imaging Guidelines board (see http://www.updig.org).
Errata: Following publication, the author requested changes to the third sentence in paragraph 6 in order to better clarify the information provided. The original sentence, as published on December 15, 2006, appears below.
The functional gaps in TMS (principally in exhibition management) will be bridged by add-on programs written by in-house applications developers.
In his keynote speech at the 2005 Museum Computer Network conference, Alexander Rose, Executive Director of the Long Now Foundation, stated poignantly that our ability to proliferate high quality images of cultural heritage materials has far out-paced our ability to sustain digital assets over time.[1] Noting a virtual sea of agreement by the wave of nods in the audience, I stifled a profound sigh of relief that this challenge was experienced so broadly. In the same room, after all, were representatives of the most esteemed art museums and cultural heritage repositories from the United States and beyond. The Center for Creative Photography was on the verge of moving from a twenty-year-old flat file database to a relational collection information system, so I had a number of burning questions on my mind. Can a collection information system also function as a digital assets management system? How can the capture of technical metadata help to sustain digital assets over time? What data elements should be embedded only in the file headers of image files vs. stored in an external database? Which elements can be automatically extracted from the file header and parsed into individual data element fields in a database so that they are searchable? Will the rights and reproductions module in the new collection information system be robust enough to fully manage the Center’s publication requests? Is it safe to begin using the JPEG2000 format?
Digital preservation anxiety, it occurred to me, was a phenomenon experienced not just by database managers and imaging professionals at universities and small museums with oft-elusive resources, but by experts from a wide array of institutions with enormously varied operating budgets. I experienced a moment of inspiration on that crisp fall morning in Boston, not, as one might suspect, for the reason that misery loves company, but because it became clear that in a room full of passionate people with active minds no one individual or institution had a perfect answer or a one-size-fits-all solution for managing and sustaining digital assets for generations to come. I was eager to get to work.
The Center embarked upon an ambitious initiative in 2004 to digitize the entire collection of 80,000 photographs over a four-year period. Coined “The Exposure Project,” the initiative is designed to create free public access to a searchable database; an engaging, interactive, multimedia tour into the archives; and a guided educational experience about photographers’ contributions to the history of the medium. Institutional home to the archives of Ansel Adams, Wynn Bullock, Harry Callahan, Louise Dahl-Wolfe, Aaron Siskind, Frederick Sommer, Edward Weston, Garry Winogrand, as well as many other classic and contemporary collections, the Center is uniquely poised to meet the needs of researchers, curators, publishers, students of photographic history, and a general public interested in creative photography of the 20th century. When the new website launches in 2009, a diverse global audience will be able to explore the Center’s collections online with unprecedented ease and functionality.
Figure 1. With the anticipated launch of a new website in 2009, interactive modules based on photographer’s lives or themes within the history of the medium will enable unprecedented access to the Center’s photograph collection and archives. The prototype, “Edward Weston in Mexico,” demonstrates some of the functionality the Center is building into its future website.
Figure 2. View of magnification tool in Interactive Gallery section of “Edward Weston in Mexico.” (See circle of magnification in upper right hand corner.)
Figure 3. This view of the Documents Section of the Web prototype depicts an enlarged section of the first page of a letter written to Edward Weston’s son Brett in 1923, along with a transcription of the letter to the left.
Initial preparations for the Exposure Project propelled the Center for Creative Photography staff and the University of Arizona Digital Library Information Systems Team to research the top rated collection information systems and reach consensus on the best match for the Center’s holdings.[2] After a year-long request for proposal process, we selected MINISIS Inc., a vendor that offers the integrated system called MINT, designed specifically to accommodate art/museum collections, archive collections, and library collections. Recognizing that a well-designed infrastructure for storage and retrieval of digital assets was critical to support the Exposure Project, the Center’s Imaging Department and the Digital Library Information Systems Team worked concurrently to design a technical infrastructure and a preservation plan with a high degree of reliability. Projections for ample storage and backup were calculated based on the combined file sizes of the Center’s master or preservation quality TIFFs and three cropped derivatives, optimized at specific sizes for access. Key staff from the systems team performed research on reliability models in digital preservation to prevent data loss.[3] Ultimately, a system architecture utilizing two mirrored servers and two backup tapes was specified to meet the objective of housing the new collection information system and backing up the Center’s digital assets.[4] UNIX servers and backup hardware were purchased and configured. The Center’s existing digital assets were migrated to the new servers in September 2006.
As Head of Digital Initiatives and Imaging, my overarching goal during the joint application development phase of implementing MINT was to determine the potential for organizing and retrieving digital assets within the collection information system. As primary copyright administrator at the Center for over a decade, it was also within my purview to brainstorm functional improvements to the MINT Rights and Reproductions module with my colleagues.
Like Franziska Frey’s persistent call for use of standardized calibration targets in her “NEDCC School for Scanning” presentations, Günter Waibel’s impassioned pleas for automating capture of technical metadata were embedded in my mind from the Museum Computer Network and other imaging conferences I had attended. RLG’s Automatic Exposure initiative [5] had begun to have an impact on some of the high end scanning manufacturers and the Center’s Better Light scanning back, for one, produced a comma-delimited string of technical metadata, viewable through Adobe PhotoShop’s File Info window. Opening image files one by one to view technical metadata, however, can be tedious, especially when working with groups of large images at staff computers with less performance capability than those configured for graphic applications. MINISIS demonstrated a high degree of flexibility in regard to customizing its MINT software, and after several round table sessions, Christopher Burcsik, CEO/Directeur Général, MINISIS, suggested there was plenty of real estate in the product for a set of technical metadata elements to live within the same system, but separately from the original object records.
Although NISO’s Technical Metadata for Digital Still Images (Z39.87) [6] was still a working draft and, potentially, a moving target, MINISIS took the leap to partially implement the standard for the Center’s first prototype. Revisiting the standard, I scrutinized every Z39.87 data element and made a selection of those that I deemed useful for such purposes as quickly referencing file attributes without opening the image, generating periodic statistical progress reports on quantities and types of image files produced, locating master scans on the server, sharing image metadata in a collaborative digital environment, facilitating migration activities, documenting file format conversions, recording target data for quality assessment and other as-of-yet-unforeseen activities over time.
By referencing a spreadsheet mapping TIFF tag numbers to the corresponding Z39.87 data elements, [7] MINISIS software engineers created code that extracts the tagged information from the file header and parses it into corresponding technical metadata fields in the collection information system imaging module. As images are individually scanned, or later evoked through MINT, the software collects the technical metadata from the file headers and saves it in an associated database table.
Once the Center’s version of MINT is fully customized, the technical metadata field names will match the Z39.87 data element field names as closely as possible. The help menu in MINT will evoke the definition for each data element directly from the NISO data dictionary “Technical Metadata for Digital Still Images.”
Figure 4. Data entry screen in development for MINT Image Reproduction module at Center for Creative Photography. The location of the master scan is visible only to Rights and Reproductions staff through password access.
Figure 5. Full prototype screen view of preservation scan linked to CCP catalog record 82:011:003. Photograph by Edward Weston, Tina, Glendale, 1921, gelatin silver print.
Figure 6. MINT Copyright screen in development for the Center and sample administrative metadata associated with photograph (81:110:003) by Edward Weston.
Figure 7. Prototype screen shot for Image Properties. Technical metadata screens in development for the Center are based on the NISO Z39.87 standard.
Figure 8. Prototype screen shot and sample Capture Metadata in the MINT Image Reproduction Module. Metadata will be automatically extracted from image file headers, where possible. Information can also be globally batched into the system, set as default, or entered at an item level.
One of the greatest challenges for the Center’s Rights and Reproduction Department in digitizing photographs over a period of dramatic technology growth was that our scanning production periodically outpaced our server capacity. Masters were consequently stored in a variety of compartmentalized locations as we raised funds or waited for the technical infrastructure to evolve. Digital imaging was not new to the Center with the advent of the Exposure Project. Methodical scanning of selected collections commenced with the founding of the Art Museum Imaging Consortium (AMICO) in 1997. At that time, scans were produced from 35mm or 4"x5" photographic intermediaries. Reference sized JPEGs were systematically created from the captures and linked to the Center’s flat file catalog database (Inmagic). The derivatives were stored on network servers and backed up to tape. Masters were stored on external hard drives and backed up to CD gold media. Over the six-year life of AMICO, approximately 10% of the collection was scanned at a quality standard specified for online research, reference prints, and for publication at no larger than ¼ page in size. Using the Windows operating system for file management and a few rudimentary data elements fields to record information about image files in the database, it was often tedious to determine the quality of scans associated with original objects in the collection and to locate the master files efficiently.
When the Center acquired a Better Light Super 6K scanning back in 2002, the practice of scanning photographic intermediaries was abandoned. Direct digital capture enabled a new quality standard to be implemented. Through publication projects, research, and technical mentoring from pre-press digital experts, master scans were consequently standardized to approximately 50MB, neutral gray balanced and captured with a beautiful, long tonal curve that rendered specific values in the shadows, midtones and highlights. Because approximately 90% of the Center’s photograph collection consists of monochromatic prints at a size of 16"x20" or smaller, this quality standard for master scans provides more than enough data to produce very high quality reproductions for publication purposes. With the release of the National Archives and Records Administration (NARA) guidelines[8] for scanning archival collections in 2004, the image capture parameters for the Center’s preservation scans were revised. Masters now vary in size from 40MB to 120MB. Targeting a one-to-one copy ratio, the image dimensions are scaled to match the original object dimensions without magnification or reduction.
Once MINT is fully implemented at the Center, the Z39.87 data elements, such as objectIdentifierValue, file size, image height, and image width, will allow Center staff to quickly retrieve, assess, and classify the scan quality of any image files associated with a given collection record. Additional image metadata can be batched, populated from the catalog record, or otherwise annotated individually. Image quality can be classified by use of a customized drop down list in the MINT imaging module. We identified five quality levels for scans: preservation, publication, presentation, research, and reference. The Help menu will display parameters for each quality level. This feature will make the process of filling publication requests more efficient. Staff will be able to access presentation quality images from the system for lectures and other purposes. In the future, we expect that some collections may come to the Center with existing scans as part of a package. The data element Image Producer enables the name of an external organization to be recorded. Whether the collection was scanned in-house, by an external organization, or both, the system will be able to manage all the images associated with a photograph and indicate the quality of each surrogate digital file through a distinct set of nested technical metadata elements.
We expect to be able to perform queries and produce reports utilizing technical image metadata as well. Customized reports utilizing the Z39.87 data element Image Producer, for instance, in conjunction with DateTimeCreated will allow periodic scan production statistics to be generated. The imaging module in MINT also contains a data element for an individual scan technician, which will allow production to be reported in finer detail. As a result of the collaboration with MINISIS, Center staff developed a new workflow for processing recent acquisitions through the collection information system. Imaging of acquisitions will now occur between the registration and cataloging processes rather than after cataloging. This change in workflow will affect the efficiency of cataloging original prints. Activities such as the application of generic nouns to the catalog record, for instance, can be performed without handling the originals. We anticipate the collection information system will also allow born digital files, such as documentation of exhibitions, special events, and institutional history, to be processed through the Archives module. Although we do not yet know the full potential of implementing the Z39.87 standard, we intend to track its usefulness over time.
The ability of MINISIS to incorporate a series of new data elements and enhanced drop down item lists for licensing and recording use of works in the collection will determine if the MINT rights and reproductions module proves robust enough to replace the Center’s current system for managing publication requests and photographic research services. Customizations are in the works and the results look very promising. The new collection information system will allow an exchange of data with our accounting system (QuickBooks Pro) such as exporting client information, publication details, and other specific data relating to reproduction of works in the collection. To generate labels, invoices, and permission letters, data elements such as artist, title, date, copyright statement, and collection credit will be populated automatically into the rights and reproductions module from existing fields in the cataloging module. Although much of the client information and copyright data elements were defined in the existing MINT Rights and Reproductions module, we identified a number of additional data elements relating to publication details,[9] and client status that will allow new ways to track specific use of the collections for statistical reporting purposes and to generate customized permission letters.
In the future, publication details recorded during the process of licensing photographs can be used to populate the literary references field in the cataloging module. We also anticipate that when the Center’s Web application is implemented, and the collections are fully searchable, our on-line publication request form will be configured to populate client contact information and publication details in the rights and reproductions module of the collection information system.
We made slight modifications to the MINT copyright information screen in the rights and reproductions module as well. For instance, we enhanced the dropdown list for the existing MINT data element field “copyright level” to record levels of permissions granted to the Center by copyright owners in more detail. In addition to the existing MINT data element “copyright owner,” we added the terms “copyright holder,” “copyright administrator,” and “copyright beneficiary” to more accurately record some of the complexities of copyright ownership or assigned management of intellectual property.
As imaging staff, students, and volunteers contribute to the proliferation of derivative images at the Center by cropping, optimizing, and making three access copies that we archive along with the preservation master TIFFs, I have watched the development of JPEG 2000 (JP2) with much interest. Although much of the post-scanning production work is automated through actions and applied to entire folders of image files at a time, it requires repetitive work and meticulous attention to detail. As one can imagine, the image file directories can become unwieldy as large collections of work are completed. For instance, the directory for a recently completed collection of nearly 2,500 Ansel Adams photographs contains a total of 10,000 master and surrogate images. With our new preservation model in place, this means there are ultimately 40,000 image files for the Adams photographs alone. In the midst of our work with MINISIS, it was serendipitously revealed to me that the imaging software MINT employs utilizes JP2 for serving a multitude of proportionally sized images in reports and data entry screens as well as on client websites. Another stroke of luck: Michael Marcellin, one of the developers of the wavelet compression used in JP2, was on the faculty at the University of Arizona. I invited him to lunch, posed some questions about managing digital assets with the aid of the NISO standard and the use of JP2 and returned to work convinced that the Center was on to something useful. Professor Marcellin kindly returned to the Center with me and demystified the format for us. JP2 has the capacity to send only the data requested by a client, which allows you to serve images of any size from a single image file. The quality is such that it pixelates visibly only under the greatest magnification. With the added insurance that the Department of Defense and the Library of Congress had adopted JP2, I became convinced that we could simplify our imaging workflow tremendously by discontinuing the process of making three specifically sized derivative copies of each preservation quality scan, and make just one cropped JP2 copy surrogate of each master scan, to serve a multitude of access purposes. I anticipate our image directories and backups will be much easier to manage with just two files for each original photograph imaged.
The challenges of maintaining an ample infrastructure to sustain high quality digital assets over time and increasing access to collections are shared by many. Limited financial resources, minimal staffing, shifting standards, evolving file formats, expensive digital asset management and collection management software, all create the undeniable need for museum professionals to be creative in the ways technology is implemented to accomplish the strategic goals of their institutions. It may not be ideal for other institutions to create hybrid systems to approximate the functionality that an off-the-shelf digital asset management system can deliver, but it seems an efficient means of achieving our goals with the resources at hand. It is not likely that a collection information system can take the place of an accounting system, but with relational configurations, data can be exchanged in ways that will create new efficiencies in managing client services and tracking collection use. By migrating twenty-five years of rich collection data to a relational database, capturing technical image metadata in a searchable format, improving the functionality of the rights and reproductions module in the new collection information system, and streamlining production of surrogate images, the Center will soon be in a position to share and preserve its collections in ways that we could previously only imagine.
4. RLG & NARA’s white paper (RLG and NARA (2005) Audit Checklist for Certifying Digital Repositories (http://www.rlg.org/en/pdfs/rlgnara-repositorieschecklist.pdf) states in section D1.5 that a repository must have an effective mechanism to prevent loss of more than 0.001% of the collection per year. Noting that the 3-copy modeling (see the footnote above) does not meet the requirement, Han used a Markov conceptual modeling tool and calculated that a 4-copy model (2 in disks and 2 in tapes) fulfills the requirement.
6. NISO's Technical Metadata for Digital Still Images has now been balloted and approved. Now a formal standard, NISO Z39.87-2006 can be found here. The accompanying NISO Metadata for Images in XML (MIX) schema, version 1.0, is available on the Library of Congress's MIX website.
7. Data maps provided by Günter Waibel, RLG Programs, OCLC.
8. For detailed recommendations on image parameters, see Technical Guidelines for Digitizing Archival Materials for Electronic Access: Creation of Production Master Files - Raster Images, page 54, on the National Archives and Records Administration website. http://www.archives.gov/preservation/technical/guidelines.pdf
9. Rights and Reproductions Staff at the Center added the following data elements to the Mint Rights Module: Publisher Name, Publication Title, Publication Author, Publication Type (pick list), Projected Publication Date (to be entered when order placed), Actual Publication Date (to be entered after publication), Publication Page number(s), Size of Reproduction in Relation to Page (predefined pick list), Print Run (predefined pick list), Distribution (predefined pick list), Languages (global list, can select more than one).
Launched in February 2006, mcn.edu is the online face of the Museum Computer Network (MCN), a nonprofit organization with 400+ “…members representing a wide range of information professionals from hundreds of museums and cultural heritage institutions in the United States and around the world.” Much more information about MCN can be found on the “About Us” pages, including membership information, by-laws, board members, and committees.
MCN sponsors an annual conference that aims to foster “the cultural aims of museums through the use of computer technologies.” The website provides conference details and public access to presentation materials from several of its past 34 conferences. (Note: Materials from MCN 2006, recently held in Pasadena, California, should be posted by the time this issue of RLG DigiNews is released.)
The website also features resource pages that provide links to relevant organizations, publications, and funding sources, as well as Special Interest Group (SIG) pages that host additional resources compiled by MCN SIGs. For example, the Standards SIG maintains an annotated list of computer-related and controlled vocabulary standards most useful for museum functions.
There are also “Members-only” pages that offer the means to update member information and a member directory searchable by name, institution, and even areas of expertise for member-to-member networking and informal support queries. The Members-only space also provides a place for distributing documents and for conducting the organization's business, such as soliciting comments on the draft of MCN Strategic Plan.
The Missing Dimension in Web-based Museum Exhibitions: Obstacles to Adding Depth to Digital Data
Author: Richard Entlich - Cornell University Library (rge1@cornell.edu)
In the late 1990s, virtual museum exhibitions based on three dimensional imagery were a growing trend. Why aren’t there more of them available today?
In May 1974, the cover of Scientific American magazine featured a three dimensional computer graphic rendition of the then recently opened Johnson Museum of Art on the Cornell University campus. The corresponding article by Cornell professor Donald P. Greenberg [1] described groundbreaking work in the use of computer graphics for generating realistic three dimensional (hereafter 3D) scenes to simulate the visual impact of a building depending on its location and the viewer’s vantage point. In the intervening decades, use of computer aided design (CAD) has become a mainstream application in architecture, planning, and many other fields, But despite the ease with which the technology can now be used to design and site a museum building, the use of 3D graphics to advance the core mission of the museum as an institution—to document and support research into the natural history of our planet, human culture, and civilization, and to educate and entertain the general public about our collective past, present, and future—has made much less progress. This FAQ provides a brief discussion about the development of 3D technology on the World Wide Web, its adoption by museums, and the obstacles that continue to keep the creation, dissemination, and management of 3D imagery via the Web from becoming a fully mainstreamed activity.
Introduction
Museums and other cultural heritage institutions have long been rooted in place because they are based on collections of physical artifacts. Though museum exhibits can and do travel, and museums routinely loan items to each other for exhibit or study, for the most part museums have relied on available technology to advertise and expand the reach of their collections through the use of surrogates. Early surrogates took the form of sketches, drawings, and eventually photographs, especially in the form of exhibit catalogs. Museums have also used television, video, laserdiscs, cd-roms, and dvds to entice audiences to visit in person, and also to share parts of the collection with those unable or unwilling to travel.
In the early days of the Internet, some museums offered digital still photos of some artifacts on gopher or ftp sites. Following the development of the World Wide Web, and particularly after graphical Web browsers started to become widely deployed, museums were quick to recognize the medium’s potential for expanding their visitor base.[2] Virtual tours and exhibitions became popular.
However, unlike libraries, museums often house collections that can’t be fully and accurately represented by traditional two dimensional camera photography, scanning, and videography.[3] Thus, there has been long-standing interest in the museum community in means to enhance the dimensionality of virtual artifacts and exhibitions.
3D capture and presentation of museum artifacts offers a number of advantages for outreach and curation. 3D more effectively telegraphs the visual impact of sculptures, dioramas, skeletons, and other artifacts than 2D surrogates. For museums confronting inadequate exhibit space and whose collections are primarily solid artifacts, 3D allows items not on display to be accurately represented in virtual form. Fragile objects, including those that should not be exposed to light for extended periods, can be made available around the clock without damage or wear and tear.
The benefits for scholarship are also significant. Availability of digital surrogates provides access to interested parties regardless of ability to travel. Quality 3D representations are more likely than 2D surrogates to be able to substitute for direct access to an artifact because they maintain correct positional relationships and allow every feature of an object to be viewed. Most elements that would be hidden or surfaces that would be obscured by other parts in a conventional photograph are revealed because a 3D scan can capture an object from every conceivable viewpoint. Some 3D technologies, such as 3DCT (Computed Tomography) permit virtual dissections or autopsies of remote objects, giving access to internal structures and makeup that would not otherwise be visible even with the object immediately at hand. Coupled with accurate metadata representing the dimensions and internal relationships of a complex shape, which can be captured as part of the 3D scanning process, a 3D virtual object can be compared to similar artifacts in terms of overall appearance as well as size, shape, color, and texture in order to aid in identification of its age, creator, or construction techniques and materials.[4]
With all these advantages, one would expect to find 3D being routinely used by museums for a variety of purposes, but it’s the exception rather than the rule to find 3D objects on a museum’s website. The main focus is on 2D digital photography. It’s not for lack of interest, but other factors have slowed down and hindered its adoption.
Brief History of 3D on the Web
One of the earliest web standards to emerge was for 3D content. VRML (Virtual Reality Modeling Language) version 1.0 (an extremely rudimentary standard) was released in 1994 and a significantly more mature version 2.0 was finalized in summer of 1996. By 1997, when VRML 2.0 was approved by ISO and renamed VRML97, a flurry of activity around 3D ensued. Many new products were introduced and web sites featuring sample applications of 3D content sprung up everywhere.
Yet today, if one does a Web search for VRML or VRML97, many of the sites returned are frozen in time, with no updates since the late 1990s. There are certainly examples of 3D content on the web, including at museum sites, but 3D has nothing like the penetration of other web-extending rich media technologies such as Flash, QuickTime movies, or PDF. Why hasn’t 3D on the Web blossomed?
Several factors have contributed to the slow uptake of 3D technology on the Web. VRML, when first introduced, was ahead of its time. Mainstream Web browsers didn’t get the capacity to routinely handle even 2D images like GIFs and JPEGs until 1996. Not only were telecommunications links of the day too slow to comfortably handle the volume of data required for 3D, but desktop PCs of the mid 1990s simply lacked the computing and graphics horsepower to manage sophisticated realtime 3D graphics. It’s no coincidence that the early corporate force behind VRML was Silicon Graphics, Inc. (SGI), a maker of high-end graphics workstations running the Unix operating system. The first version of SGI’s 3D Web renderer (which they called Cosmo Player) was only for IRIX, SGI’s version of Unix.
After VRML97 became a standard, the Cosmo Software division of SGI released a new version of Cosmo Player for other platforms and made a major push to distribute it. Unfortunately, the prospects for VRML were dealt a crushing blow when SGI (the overall company) ran into serious business problems just as momentum for VRML was building. Lacking the resources to shepherd the new technology through a money-losing development and marketing phase, SGI tried unsuccessfully to find an appropriate corporate buyer for the Cosmo Software division, and it eventually wound up in the hands of computing giant Computer Associates, which abandoned it. Some early proponents of VRML declared it dead. (Note: Cosmo Player, though not updated in years, is still available. See table 1 for details.)[5]
3D on the Web then went through an extended period of reorganization. The VRML Consortium, the body with official responsibility for the VRML standard, renamed itself the Web3D Consortium late in 1998. It then took over three years for a new standard, X3D (eXtensible 3D, which is based on XML, the eXtensible Markup Language) to be developed. X3D became an ISO standard in mid-2004. Only recently has 3D on the Web regained significant momentum, with the emergence of tools for developers (e.g., Xj3D, a Java based toolkit and browser, and OpenVRML, a set of libraries to add VRML and X3D capability to applications) and for end users, such as FreeWRL, an open source VRML and X3D browser primarily for the Linux, Unix, and Mac OS X platforms.
Numerous consumer oriented software packages for accessing 3D on the Web have come and gone over the past decade. ExtremeTech, a Web publishing arm of PC Magazine, published a roundup of currently available Web browser plug-ins for VRML and X3D in October 2006, the first we’ve seen in many years, and possibly indicative of renewed interest. NIST (the US National Institute of Standards and Technology) maintains a VRML Plugin and Browser Detector that probes Web browsers for installed VRML and X3D plugins and also provides a wealth of information and external links related to VRML and X3D plugins.
3D and Museums
The museum community was an early adopter of the Web as a vehicle for communicating with the public. Some museums had websites up by early 1994.[6] The Museums and the Web conference, put on by Archives & Museum Informatics, was established in 1997 and has been held annually for a decade. The first paper from that conference to mention use of 3D technology in its title was presented in 1999, and twenty more papers, workshops, and demonstrations with similarly prominent mentions of 3D have been featured since then. Interest, enthusiasm and a willingness to experiment have all been available in abundance.
Despite the healthy level of interest, however, the environment for deployment of 3D on the Web by museums has been a frustrating one. The turmoil over development of the VRML and X3D standards and the tools to develop, disseminate, and view content is only one of the obstacles museums have had to face. The failure of VRML and X3D to become mainstreamed has resulted in heavy fragmentation of the 3D market. Numerous proprietary products and a competing standard, U3D (an Ecma standard that is supported in Acrobat Reader v.7.0) are available. Even the basic terminology is fragmented. For example, finding all the Museums and the Web titles that reference 3D technologies requires searching on ‘3D,’ ‘3-D,’ and ‘three dimensional.’
Adding to the confusion, not all Web objects commonly labeled as 3D offer the same degree of functionality. VRML97, X3D, and other true 3D technologies allow the construction of scenes that can be freely explored from arbitrary vantage points, though not necessarily with a particularly high level of graphical realism. However, both museums and the users of their websites are more likely to be familiar with simulated 3D technologies, especially QuickTime VR, but also Shockwave3D, and Flash animations. These may allow the user to rotate an object or pan through a scene, but lack an internal 3D model and are typically constructed from a series of 2D photos of an object or location that have been digitally stitched together for convenient viewing.
With no truly dominant technology, museums have chosen a variety of approaches for making true 3D and simulated 3D content available via the Web. This requires prospective consumers of 3D offerings to install and maintain a large group of potentially incompatible plug-ins and helper applications. Table 1 shows just a small sample of the range of technologies in use by museums and research institutions around the world to disseminate 3D content. The examples include both simulated and true 3D technologies.
Table 1. Examples of 3D-based exhibits and artifacts on the Web, deployed using a variety of technologies.
The current status quo also leaves museums that have decided to offer 3D content on the Web with some tough choices. These are well summarized in a 2002 paper by the creators of the Inuit 3D exhibit at the Canadian Museum of Civilization:
... an interactive virtual Web exhibition requires a careful selection and integration of 2D, 3D and audio-visual Web application tools and related plug-ins. Constraints such as bandwidth, browser differences, file sizes, plug-in requirements and 3D model compression lurk half-submerged like alligators in a swamp, ready to take a bite out of plans to create virtual exhibitions that are only a few bricks short of looking like the real thing. One must consider trade-offs between image quality and download time, the “download tolerance” of viewers and whether the goal of the exhibition is to provide entertainment, edutainment or an educational experience.[7]
Management of 3D Digital Assets in Museums
The lack of standardization and the failure of any single 3D Web technology to control a large segment of the market is also hard on those responsible for maintaining 3D digital assets. In addition to the problems already described, there is a trickle-down impact on the museum systems that have to manage the non-standard digital content.
A digital asset management system (DAMS), as discussed elsewhere in this issue, offers the potential to flexibly repurpose virtualized collections and tightly control who gets access to what version of what content. A DAMS can store and describe essentially anything that can be represented digitally, but the level of support can vary substantially depending on the file type and format. For example, a DAMS may provide a metadata scheme that is thoughtfully geared to capturing the qualities of still image files or text documents, but that may be much less adequate for providing usable search, retrieval, and long-term management functionality for more complex digital objects.[8]
Also, the ability to create a thumbnail representation of a file, preview it, alter its quality or size, and repurpose it through file format or file type transformation is likely to be better supported for some kinds of assets than for others. Creating thumbnails and previews of still images, creating derivatives of different size and quality, and converting among different still image formats are routine operations that are widely supported by the average DAMS. For commonplace rich media types, such as audio and video, many DAMS already offer specially targeted features, such as the ability to identify key frames to use for searching and representation, the ability to preview file contents, some support for derivative creation (e.g., changing the bit depth of an audio file or the frame rate of a video file) and transcoding of the files.
This functionality may either be built in to the DAMS, be an optional add-on offered by the product vendor, or available through integration with a third party product. A popular partner for DAMS vendors is Telestream, which makes the FlipFactory line of products. For example, the FlipFactory DAM product offers “multiformat digital video and audio ingest, metadata extraction, keyframe extraction for storyboarding, delivery of extracted metadata to the DAM system, automated proxy generation, transcoding between formats, and repurposing to Web, mobile, etc.” Such features extend the ability of a DAMS to manage audio and video content.
3D content within a DAMS shares some requirements of other rich media types, but also presents some unique challenges. Like audio and video, 3D files cannot be accurately represented by a single thumbnail-type surrogate. A virtual tour of a museum rendered in VRML or X3D might encompass views of every gallery and many individual artifacts. What are the “key frames” of such a file? What kind of preview should be created that will adequately suggest the potential uses of the resource? What kinds of derivatives or other formats can be created? We were unable to find any DAMS that specifically provided support for true 3D objects beyond basic storage and metadata.[9] If previewing is supported at all, it’s through use a specified external program. This low level of support is a reflection and consequence of the unsettled state of 3D on the Web.
Conclusion
3D digital surrogates of museum exhibits and artifacts have a great deal of potential for providing enhanced virtual access to scholars and the general public alike. Though 3D computer graphics has a long history and has become a mature industry in some sectors, development of 3D for Web delivery has been hampered by a lack of well-supported standards, insufficient interest from leading software companies, and the absence of stable open-source programs for creating and rendering content. Furthermore, delivering high-quality 3D via the Web requires state-of-the-art networks, computers, and graphics processors. Finally, DAMS lack basic functionality for managing 3D digital content, so that museums that have created such assets either cannot integrate them with existing management systems or must tolerate the absence of features that are available for most other types of digital content.
However, the future for 3D on the Web is looking better. X3D has become an ISO standard and some open source developer and end user tools that support it are starting to emerge. Broadband networks are gradually increasing in capacity as well as degree of market penetration. The release of Microsoft’s Windows Vista operating system is widely expected to precipitate a new round of desktop computer upgrades, and the high-end version of Vista’s graphical user interface, dubbed Aero Glass 3D, requires an accelerated 3D graphics card to run. This may help move high end accelerated 3D graphics from the specialized domain of hardcore gamers into the mainstream.
As with any technology, it’s hard to predict when events will propel 3D on the Web to a critical mass of interest that will break the current chicken and egg logjam. More than five years ago, a feature article in this publication by Tony Gill [10] included the statement “Most industry pundits agree that the Web 3D revolution is now long overdue.” Given its history of confounding the pundits, all it seems prudent to say now is that recent progress on Web 3D is promising enough to merit keeping an eye out for new developments.
Acknowledgements: This report was supported in part by a grant from the Institute of Museum and Library Services for the KMODDL (Kinematic Models for Design Digital Library) Project. A white paper with more in-depth coverage of the uses of 3D digital content by cultural heritage organizations and the digital preservation issues accompanying their use will be published in Summer 2007 under the auspices of the same grant. Thanks to Ellie Buckley and Nancy McGovern for work on KMODDL that contributed to this report, as well as to KMODDL principal investigator John Saylor and project manager Kizer Walker. I would also like to thank guest editor Günter Waibel for his assistance and advice.
Notes
1.Greenberg, Donald P., “Computer Graphics in Architecture,” Scientific American, v.230, May 1974, pp.98-106.
2. Results from a survey of nearly three dozen art museums published in 2000 found the single most common reason for mounting virtual exhibits, at 44%, was “to encourage individuals to visit the physical exhibit,” while only 26% cited “so all individuals who are not able to visit the museum can view materials remotely.” See Nystrom, Margaret J., “Virtual Art Museum Exhibits on the World Wide Web: A Content and Survey Analysis,” Master’s paper, School of Information and Library Science, University of North Carolina at Chapel Hill, November 2000.
3. Arguments have also been made that some digitized library materials should be scanned in 3D to preserve the look and feel of materials that are “warped and crinkled due to decay, neglect, accident and the passing of time.” See Brown, Michael S., and W. Brent Seales, “Beyond 2D Images: Effective 3D Imaging for Library Materials,” Proceedings of the fifth ACM conference on Digital libraries, p.27-36, June 2-7, 2000, San Antonio, Texas.
4. For extended discussions of the value of 3D digital surrogates for remote research on artifacts, see Subodh Kumar, Dean Snyder, Donald Duncan, Jonathan Cohen, and Jerry Cooper, “Digital Preservation of Ancient Cuneiform Tablets Using 3D-Scanning,” Proceedings of Fourth International Conference on 3D Digital Imaging and Modeling. pp. 326-333. October 2003 [section 1.1]; and Clark, Jeffrey T., Brian M. Slator, William Perrizo, James E. Landrum, III, Richard Frovarp, Aaron Bergstrom, Sanjay Ramaswamy and William Jockheck, “Digital Archive Network for Anthropology,” JoDI: Journal of Digital Information, v.2 no.4, 2002 [section 2].
6. Bowen, Jonathan P., Jim Bennett and James Johnson, “Virtual Visits to Virtual Museums,” in J. Trant and D. Bearman (eds.) Museums and the Web 1998: Proceedings, Toronto: Archives and Museum Informatics, 1998.
7. Corcoran, Frank, Jeffrey Demaine, Michel Picard, Louis-Guy Dicaire and John Taylor, “INUIT3D: An Interactive Virtual 3D Web Exhibition,” in J. Trant and D. Bearman (eds.) Museums and the Web 2002: Proceedings, Toronto: Archives and Museum Informatics, 2002.
8. Some work has been done on developing metadata standards for 3D on the web. See, for example, Lewis, Elise, Samantha K. Hastings, and Cathy Nelson Hartman, “Recommendations for Metadata Standards for 3D Images on the Web,” Final Program and Proceedings of IS&T's 2004 Archiving Conference, San Antonio, Texas, April 20-23, 2004, pp.187-190.
The Museums Copyright Group and King’s College London will host this one-day forum that intends to explore and debate the relationship between culture and commerce in cultural heritage institutions. Participants from museum, library, archive, and education communities along with those from the creative industries, media, legal, commerce, and other business sectors are encouraged to attend.
This conference aims to “create collaborative, cross-departmental, cross-community approaches to look at the impact that the digital environment has on library collections, access to resources, and organizations.” Presentations and panel sessions will include hands-on and interactive formats to facilitate dialog among participants from different fields both within and outside the library. Look for sessions on topics including: collaborative e-resources management, collection development of e-resources, institutional repositories, and others.
The 2007 WebWise Conference on Libraries and Museums in the Digital World will be co-hosted by IMLS, the J. Paul Getty Trust, and OCLC Online Computer Library Center. This year’s theme is “Stewardship in the Digital Age: Managing Museum and Library Collections for Preservation and Use.” Topics to be addressed include emerging practices for the preservation of digitized text, issues facing “born-digital” art, the “digital readiness” of museums and libraries, and technological strategies for improving collections management and care. (Note: at the time of the release of this issue of RLG DigiNews the link to the conference was not yet made public, however it should go live within a few days.)
The International Association for Social Science Information Service and Technology has issued a call for papers for its 33rd annual conference, themed: Building Global Knowledge Communities with Open Data. The conference will include pre-conference workshops, platform sessions, and posters followed by a weekend of optional activities in the Montreal area. Proposals will be accepted through January 16, 2007.
This annual conference targets professionals from cultural heritage, IT, media, government, and media research sectors. The focus of the workshops and training sessions will be centered around electronic imaging and the visual arts.
The Digital Preservation Coalition has announced a call for entries for the third Digital Preservation Award. Applicants must demonstrate “how a completed project focusing on preserving digital materials (whether “born digital” or digitized copies) provides leadership and advancement in digital preservation which will benefit the UK.” Applicants from outside the UK are welcomed, but the project must demonstrate benefit to the UK. The deadline for applications is March 31, 2007.
The North Carolina Archives and State Library of North Carolina has announced the launch of its Web Site Archive. Using the Internet Archive’s Archive-It technology, the project has developed a tool that captures, preserves, and provides easy access to both the content and look-and-feel of state government Web pages.
The International Journal of Digital Libraries has announced a call for papers for a special topic issue on “Very Large Digital Libraries.” The special issue will include research reports, case studies, survey articles, and implementation experiences emphasizing techniques for the creation, maintenance, and usage of large-scale digital libraries.
University of Virginia Library has joined Harvard, Oxford, the New York Public Library, the University of Michigan, and others in the Google Books project. Thousands of books from the Library’s American history, literature, and humanities collections will be digitized and made searchable within Google’s Book Search product.
A new report commissioned by Wesleyan University and the National Institute for Technology and Liberal Education (NITLE) was recently released. The report, “Using Digital Images in Teaching and Learning,” focuses on the implications of the widespread use of images in digital formats, especially issues concerning supply, support, and infrastructure for using digital image resources such as quality of image resources, image functionality, management, deployment, and requisite skills for usage. The study is based on four hundred survey responses and three hundred individual interviews with faculty and staff at 33 colleges and universities.
The next issue of RLG DigiNews, volume 11, number 1 (February 2007) will mark ten full calendar years of RLG DigiNews publication, a fantastic collaborative adventure between RLG and the Cornell University Library.
The first issue launched in April 1997 after a series of conversations between Nancy Elkington (RLG) Anne R. Kenney (Cornell) and Robin Dale (RLG). Over the years, it has evolved from a publication containing information about digitization projects and practices to one that covers both digitization and digital preservation, as well as the best practices, new initiatives, developing standards, and continuing education opportunities related to both topics. Readership has grown from a few hundred readers to thousands of people from more than 80 countries and consistently ranks as one of the most-visited information resources on the RLG website (link to www.rlg.org).
The past ten years have also brought evolution to the organizations involved in this publication. Anne Kenney, the longstanding leader of the RLG DigiNews team at Cornell has moved from the Associate Director of Conservation and Preservation and most recently became Cornell’s Senior Associate University Librarian. And as most people know, RLG recently combined with OCLC, leveraging the strengths of both organizations and placing RLG DigiNews into the capable hands of the newly formed OCLC Office of Programs and Research. With these great changes come new opportunities and areas of interest for both organizations and so it is with both excitement and not a small bit of melancholy that we announce that the February 2007 will be the last issue collaboratively published with Cornell.
As with any big change, we wanted to alert readers in advance. We also want to acknowledge the strong readership and support of this publication. In future months, DigiNews will be published directly by the OCLC Office of Programs and Research. Your ideas and suggestions on how we can take best advantage of this new opportunity would be welcome.
RLG DigiNews (ISSN 1093-5371) is a Web-based newsletter conceived by the RLG preservation community and developed to serve a broad readership around the world. It is produced by staff in the Department of Research, Cornell University Library, in consultation with RLG Programs, OCLC Office of Programs and Research, and is published six times a year at www.rlg.org.
Materials in RLG DigiNews are subject to copyright and other proprietary rights. Permission is hereby given to use material found here for research purposes or private study. When citing RLG DigiNews, include the article title and author referenced plus “RLG DigiNews”. Any uses other than for research or private study require written permission from RLG and/or the author of the article. To receive this, and prior to using RLG DigiNews contents in any presentations or materials you share with others, please contact Robert Bollander (bolander@oclc.org ), OCLC Office of Programs and Research.
Please send comments and questions about this or other issues to the RLG DigiNews editors.
Co-Editors: Anne R. Kenney and Nancy Y. McGovern; Associate Editor: Robin Dale (RLG Programs, OCLC); FAQ Editor: Richard Entlich; Contributor &Production: Ellie Buckley; Advisor: Peter Hirtle.
All links in this issue were confirmed accurate as of December 14, 2006.
Printable Version
it emerged as the primary reason for digitization with a solid 56% in the 2005 survey.
The last point may bear a little more scrutiny: the discussions around digital repository certification [
The Met Images project, in planning since early 2002 and under discussion long before that, is one of the most expensive non-construction projects ever undertaken by the Met. The three-year system implementation will, barring unexpected delays, be accomplished in a shorter period than the planning and discovery that led to the approval of the project by the museum’s executive staff and Board, reversing the more common ratio of planning-to-project years. The lengthy planning period was due, in part, to the Met’s commitment to asking and answering a series of questions not always built into the development process for digital asset management activities in the not-for-profit world. Our questions included quantitative ones such as: “How big is the collection of assets?” and “How much storage will be required?” We asked questions about proprietorship and access: “Who uses the assets?” and “Who will manage them?” We wondered about the quality and value of the assets: “Should images that are made available to the public be color corrected?” and “If the object’s descriptive record has not been reviewed, may it be distributed along with the asset?” and finally, “Are existing descriptions adequate to support successful searching at all?” We also asked a number of questions that forced the Met’s staff and executives to think through fundamental intellectual property policy positions: “Who decides who may use the assets, both inside and outside of the museum?” and “Is the Met’s goal to profit from the licensing of images, or to support an educational mandate for broad distribution?” All of these questions needed to be considered in light of a process that would, inevitably, seek to automate the answers. Processes and policies that had heretofore been entrusted to individuals within the organization would need to be formalized so that a system might manage them; decisions that had been made on an ad hoc basis now needed to be seen as patterns that formed policies. And as we found answers to our questions, the scope of the project inevitably grew. 
well-organized, fully-populated image database, for example—none of the organizations we talked to had achieved a complete end-to-end solution. Only a few other institutions seemed even to aspire to such an integrated system, perhaps because in most organizations no one department or individual bore responsibility for all of the functions of a system such as the one we envisioned. Nonetheless, at the end of about a half-year of research, the Met team had determined that such a tool set was, indeed, our goal. We knew from the start that the project would be expensive and thus would require the allocation of capital funding by the museum’s Board of Directors and Finance Committee. A request to the Board requires—at the Met as at most museums—a detailed financial request accompanied by a “budget justification.” The Met Images team considered several budget justification strategies in the course of our needs assessment planning and research. We began by crafting a budget justification based on a return-on-investment rationale: the system costs would be paid for through the revenue and efficiencies that the new tools would generate. 
The Met’s Board responded to the initial project proposal by asking two questions that significantly reshaped the project plan and justification. First, they asked the team to consider whether the project might be staged in phases, allowing for a review of the requirements and goals at the end of each year. Second, they wondered whether return on investment or profit were in fact the right drivers for this project. They suggested that the project’s goal might, instead, be about essential investments in the museum’s assets and infrastructure that would support core mission objectives: to research, document, and educate. With this guidance in place, the project was re-structured into three phases: the first primarily about the acquisition and deployment of the repository; the second about the broadening of access to staff throughout the museum, integration with other museum applications, and the implementation of more sophisticated security structures; and the third about enabling access to assets through a self-serve, Web-based front end. Despite the shift from a return-on-investment to a mission-driven rationale, the project retained its inventory and cataloguing elements, building into each phase activities intended to develop additional inventory and enhance cataloguing. The decision to maintain this emphasis on content development has helped the project to achieve widespread support throughout the institution. 
A system such as the one we proposed to build would contain tools that would support access to and distribution of museum assets at a far lower cost than had been possible in the previous system—which combined a high personnel cost (human beings had to review, research, and respond to each individual request for access to assets) and a high material cost (analog images, in particular, had a limited “shelf life” and were easily lost, damaged, or worn out). When it became clear that the new tools would significantly reduce “transaction” costs associated with fulfilling individual requests for images, we asked ourselves whether it was appropriate to continue to offer the same fees that we had previously—charges that had been largely based on recouping the museum’s costs for fulfillment. The group felt strongly that for scholars, a class of users that the museum seeks actively to support, any savings realized by the implementation of a system that would reduce costs should be passed on. 
began a discussion with staff in a number of departments, including the libraries, education, and curatorial—about meaningful ways to collect subject terms, or keywords for retrieval. The group examined options ranging from engagement of permanent staff for keywording, to data mining of existing content. Every solution appeared inadequate; some were also expensive. Inspired by the work done at the Fine Arts Museums of San Francisco to deploy volunteers with index cards to describe works of art in “lay” terms, the group eventually turned to a plan that anticipated the blossoming of interest in “social tagging”—the employment of Web users to provide keywords (or tags), to label content. The Met’s internal discussions led to a broader discussion within the museum community at large and ultimately to the formation of the 
uniqueness, image quality, cataloguing quality, replacement costs, and centrality to the institution’s mission. Ideally, inventories should probably be taken on a recurring basis and concurrent with efforts to load identified collections to the digital asset management system. Though the inventory process can be labor intensive, it is an essential stage in minimizing redundant activity, identifying at-risk collections, and setting priorities for digitization, cataloguing, and ingest into the centralized digital asset management system. At the Met, we have created an inventory questionnaire to elicit information about “local” collections—the uniqueness of the materials, the size of the collection, the characteristics and format of the files, the extent of cataloguing, and the frequency with which staff members access the collection items. This inventory work has already served to identify departments managing valuable content with institutional or long-term value and also departments unaware that their material is derivative content duplicated elsewhere in a higher-resolution or better-catalogued version. 
These varied and sometimes competing management responsibilities did not readily fit into the existing structures of the Metropolitan Museum. For the project to succeed, management of both the images and the image repository needed to be distributed among staff from different areas of the museum, and new working relationships had to be developed between areas of the organization that had not collaborated closely before. The museum’s senior administration did consider whether any single department might head this effort but decided instead to create a project team, headed by the Director’s Office and Information Systems and Technology and advised by members of the Image Library, Photo Studio, Counsel’s Office, and Finance Department—collectively, the Met Images “Lead Team”—to manage the Met Images initiative while in its implementation phase. 
However, the studio was forced to engage with a series of questions not previously considered as preparation for the Met Images implementation moved forward. As the museum’s staff at large embraced the concept of a searchable, centralized repository of images, they began to wonder about the relationship of the repository to new photography ordered in support of museum programs such as exhibitions and publications. Often produced on a rush basis, such programs would be well served by up-to-date information about the status of photography orders and, better served still, by the ability to access and download new material as soon as it was ready. The existing photography workflow needed to be overhauled in order to integrate it fully with the digital asset management system and tools. This project, related but separate, constituted a major initiative in itself, and because many of the same staff members who were already working on Met Images would be involved, the work has been scheduled to occur during the second phase of Met Images.
In each of these areas, museums should look to existing and developing standards. Creating our metadata schemas based on these standards and creating crosswalks to them will allow us to better communicate with one another. Metadata is also an area where museums should proceed cautiously. DAM vendors and implementers are relatively new to the cultural heritage space. Cataloguing information about assets related to works of art is different from cataloguing information about sneakers or pickup trucks. The permanent and archival nature of our digital asset collections may require different functionality and data structures than those offered by most large commercial digital asset management vendors. The full relational content of our catalogue records from our collections management systems is not easily searchable in the flat file context of many leading systems. Nonetheless, as DAM vendors come to understand the requirements of our community better, we believe that tools will be developed that are more appropriate to the needs of museums, libraries, and archives. 
Although the management tasks are complex and sometimes unfamiliar, successful creation and adoption of them—paired with new technologies and platforms for communication of information about our collections—will expand the reach of our institution in a profound way. This much is probably true for most of the museums and cultural heritage institutions now considering the implementation of a digital asset management system. However, for each of these institutions, decisions about how to implement a digital asset management system, who will do the work, when to begin, and what assets to manage will certainly be based on informed choices at the institutional level. The decisions will, in the end, shape each organization’s digital asset management strategy into something that is distinctly its own, based on a process of institutional self-discovery that is likely to give birth to new alliances and collaborations, build new skills, and produce some outcomes that may reshape institutional policy and procedures in unexpected and fundamental ways. 
The functionalities provided by these interlocking systems should be considered as core elements in a comprehensive suite of digital asset management tools and applications.
Each master file is saved in the RAW image format at 16 bits per channel. All subsequent image editing is done on a full resolution derivative file in TIFF format with edits preserved in layers. Layers keep a saved history of any file changes, so files can be rolled back to their original states whenever necessary. The TIFF file is used to generate a third derivative file, our so-called “parent file,” with flattened layers. The parent file is used to generate other versions for specific output requirements. All three files (master, layered parent, and flattened parent) are saved and managed by Portfolio.
Additionally, the EDAM can provide a home for the multitude of research image files that are individually parked on desktops and servers throughout the Gallery without the prerequisites of descriptive cataloging and broad accessibility.
Digital preservation anxiety, it occurred to me, was a phenomenon experienced not just by database managers and imaging professionals at universities and small museums with oft-elusive resources, but by experts from a wide array of institutions with enormously varied operating budgets. I experienced a moment of inspiration on that crisp fall morning in Boston, not, as one might suspect, for the reason that misery loves company, but because it became clear that in a room full of passionate people with active minds no one individual or institution had a perfect answer or a one-size-fits-all solution for managing and sustaining digital assets for generations to come. I was eager to get to work.
Although NISO’s 
Methodical scanning of selected collections commenced with the founding of the Art Museum Imaging Consortium (
The Help menu will display parameters for each quality level. This feature will make the process of filling publication requests more efficient. Staff will be able to access presentation quality images from the system for lectures and other purposes. In the future, we expect that some collections may come to the Center with existing scans as part of a package. The data element Image Producer enables the name of an external organization to be recorded. Whether the collection was scanned in-house, by an external organization, or both, the system will be able to manage all the images associated with a photograph and indicate the quality of each surrogate digital file through a distinct set of nested technical metadata elements.
In the future, publication details recorded during the process of licensing photographs can be used to populate the literary references field in the cataloging module. We also anticipate that when the Center’s Web application is implemented, and the collections are fully searchable, our on-line publication request form will be configured to populate client contact information and publication details in the rights and reproductions module of the collection information system. 
The challenges of maintaining an ample infrastructure to sustain high quality digital assets over time and increasing access to collections are shared by many. Limited financial resources, minimal staffing, shifting standards, evolving file formats, expensive digital asset management and collection management software, all create the undeniable need for museum professionals to be creative in the ways technology is implemented to accomplish the strategic goals of their institutions. It may not be ideal for other institutions to create hybrid systems to approximate the functionality that an off-the-shelf digital asset management system can deliver, but it seems an efficient means of achieving our goals with the resources at hand. It is not likely that a collection information system can take the place of an accounting system, but with relational configurations, data can be exchanged in ways that will create new efficiencies in managing client services and tracking collection use. By migrating twenty-five years of rich collection data to a relational database, capturing technical image metadata in a searchable format, improving the functionality of the rights and reproductions module in the new collection information system, and streamlining production of surrogate images, the Center will soon be in a position to share and preserve its collections in ways that we could previously only imagine.
In the early days of the Internet, some museums offered digital still photos of some artifacts on gopher or ftp sites. Following the development of the World Wide Web, and particularly after graphical Web browsers started to become widely deployed, museums were quick to recognize the medium’s potential for expanding their visitor base.[
One of the earliest web standards to emerge was for 3D content. VRML (Virtual Reality Modeling Language) version 1.0 (an extremely rudimentary standard) was released in 1994 and a significantly more mature version 2.0 was finalized in summer of 1996. By 1997, when VRML 2.0 was approved by ISO and renamed VRML97, a flurry of activity around 3D ensued. Many new products were introduced and web sites featuring sample applications of 3D content sprung up everywhere.
After VRML97 became a standard, the Cosmo Software division of SGI released a new version of Cosmo Player for other platforms and made a major push to distribute it. Unfortunately, the prospects for VRML were dealt a crushing blow when SGI (the overall company) ran into serious business problems just as momentum for VRML was building. Lacking the resources to shepherd the new technology through a money-losing development and marketing phase, SGI tried unsuccessfully to find an appropriate corporate buyer for the Cosmo Software division, and it eventually wound up in the hands of computing giant Computer Associates, which 
3D on the Web then went through an extended period of reorganization. The VRML Consortium, the body with official responsibility for the VRML standard, renamed itself the 
Adding to the confusion, not all Web objects commonly labeled as 3D offer the same degree of functionality. VRML97, X3D, and other true 3D technologies allow the construction of scenes that can be freely explored from arbitrary vantage points, though not necessarily with a particularly high level of graphical realism. However, both museums and the users of their websites are more likely to be familiar with simulated 3D technologies, especially QuickTime VR, but also Shockwave3D, and Flash animations. These may allow the user to rotate an object or pan through a scene, but lack an internal 3D model and are typically constructed from a series of 2D photos of an object or location that have been digitally stitched together for convenient viewing.
A digital asset management system (DAMS), as discussed elsewhere in this issue, offers the potential to flexibly repurpose virtualized collections and tightly control who gets access to what version of what content. A DAMS can store and describe essentially anything that can be represented digitally, but the level of support can vary substantially depending on the file type and format. For example, a DAMS may provide a metadata scheme that is thoughtfully geared to capturing the qualities of still image files or text documents, but that may be much less adequate for providing usable search, retrieval, and long-term management functionality for more complex digital objects.[
However, the future for 3D on the Web is looking better. X3D has become an ISO standard and some open source developer and end user tools that support it are starting to emerge. Broadband networks are gradually increasing in capacity as well as degree of market penetration. The release of Microsoft’s Windows Vista operating system is widely expected to precipitate a new round of desktop computer upgrades, and the high-end version of Vista’s graphical user interface, dubbed Aero Glass 3D, requires an accelerated 3D graphics card to run. This may help move high end accelerated 3D graphics from the specialized domain of hardcore gamers into the mainstream.