The Power of Partnering: The Cooperative Creation of Digital Collections

Kathleen Foulke, Nancy Milnor*, Melissa Watterworth** and Thomas Wilsted**

Mystic Seaport, Mystic, CT 06355, USA

Email: kathleen.foulke@mysticseaport.org; Web: http://www.mysticseaport.org/

*The Connecticut Historical Society Museum, Hartford, CT 06105, USA

Email: nancy_milnor@chs.org; Web: http://www.chs.org/

**Thomas J. Dodd Research Center, University of Connecticut, Storrs, CT 06269-1205, USA

Email: melissa.watterworth@uconn.edu, tom.wilsted@uconn.edu; Web: http://www.lib.uconn.edu/online/research/speclib/ASC/

Web: http://www.cthistoryonline.org

Abstract

The use of consortia and partnerships has increased significantly over the past decade as a result of the increasing complexity of developing projects in a digital world. Funding agencies have acknowledged this by directing support to multi-institutional projects. Some agencies have gone so far as to establish categories that require multi-institutional applicants. Partnerships of disparate institutions provide opportunities for learning and growth. The paper describes one such partnership, Connecticut History Online, and investigates the value and significance that partnering plays in creating a successful digital product.

1 Connecticut History Online: phases 1 and II

1.1 Project beginnings

Connecticut History Online

(CHO) began as a partnership of three institutions seeking joint funding from the Institute of Museum and Library Services (IMLS) under IMLS's National Leadership Programs for Libraries and Museums. The consortium proposed the development of CHO, a database of 15,000 historical images drawn from the collections of the partners in 1998. Connecticut Historical Society Museum and Mystic Seaport are private institutions, while the Thomas J. Dodd Research Center is a division of the University of Connecticut, a state-funded university. The idea of partnering was initially driven by the three institutions' interest in seeking funding to catalog and digitize items from their collections and an IMLS requirement that this type of grant was only available to partners from the museum and library communities.

The project combined the museums' understanding of the power of images to promote history and culture and their experience in working with the education community; the libraries brought strengths in the use of technology for improving intellectual access to and organization of collections. All three institutions contained strong image collections, but the strengths of the collections were in different areas; thus the project also combined the best images from three important Connecticut collections with different subject strengths.

From the beginning, project partners focused their efforts on addressing the needs of a core audience: middle and high school teachers and students. Curriculum standards today at both the national and state levels require the integration of primary source materials to engage students in learning experiences rooted in historical themes, promote critical and reflective thinking, and build reading and information literacy skills (Riley 1999, Wineburg 2001, National Council for the Social Studies 1994). The project saw this user group as another partner and worked early on to bring teacher input into planning and development. The project established an education committee as well as teacher advisory groups, and held focus groups to gather responses on content, Web site design, and usability. The resulting product was significantly changed and enhanced through teacher input.

The partners received a second grant for the Phase Two proposal to IMLS in June 2002. The second grant supported the addition of two new partners, the Connecticut State Library and the New Haven Colony Historical Society. The goals of the second grant included testing methods for adding new partners of varying sizes, and adding new types of collection material including maps, manuscripts, serials, oral histories and multi-page documents. In addition, during the second phase, the partners would continue to develop programs to support middle and high school teachers and students, the core audience for CHO. Finally, the partners would test new technology and investigate options and opportunities for establishing long-term support for CHO. A new Web site containing both existing and new collections will debut in the fall of 2004.

1.2 Partnership as a core value: diversity of missions, strengths, collections

1.2.1 Commitment to common goals

Although the initial partnership was based on obtaining financial support for a common project, the partners discovered the value of their individual strengths during the initial phases of grant writing. The combination of developing and integrating collections, establishing an educational program, and implementing a state-of-the-art technical system that supported the delivery of online resources required each institution to use the knowledge and talents of different members of their staff. Top staff at all three institutions devoted their knowledge and talents to creating the initial grant. The process of creation highlighted the complementary talents and this initial introduction carried over to the development of the project once funding was assured.

CHO broke new ground as a cooperative project in a geographical region known for rugged independence. As the project evolved, the creation of this digital collection through the partnering of a library, a historical society, and a museum, institutions with greatly differing administrative structures, proved that a partnership of diverse institutions at geographic distance from each other could succeed. The CHO project, as a national model of cooperation, helped break down barriers within the state and encouraged other institutions to consider cooperation as a means of achieving joint goals.

Making the partnership work and seeing the project through to fruition was not without difficulty, as the three institutions involved had different cultures and different ways of approaching the project. However, all involved were committed to enhancing access to and use of their collections. At the end of phase I a real common ground of pride in the product, the satisfaction of difficulties overcome and the sense of a job well done had forged the representatives of these three institutions into a true partnership. The partners were sufficiently satisfied with the outcome that they decided to ask two more institutions to join the consortium and submit a grant request to IMLS for a second phase of the project.

1.2.2 Partner representation

CHO operated as a partnership of equals from the initial grant writing to the establishment of the committees, to decisions about technical standards, to the development of educational programs. The lead institutional role has rotated among the partners. The Connecticut Historical Society was the initial grantee for CHO. It drafted the first grant, received and managed the funds, provided a project director and housed the project coordinator. The grant for the second phase of CHO was developed and managed by the Thomas J. Dodd Research Center, and Mystic Seaport submitted a third CHO grant proposal on behalf of the consortium in 2004.

To meet IMLS requirements, each partner signed an agreement indicating the tasks they would complete as part of their grant responsibility. The grant enabled the partnership to create what is effectively a new entity, Connecticut History Online, originally an amalgam of three and, now, five partners. In phase I, each institution agreed to provide an equal number of images and received equal funding to outsource digitization and create metadata to describe the images. During phase II, new partners are providing more materials than the initial three. The three original partners focused on new types of material and devoted more time to developing new technologies to deliver this material.

Clearly some institutions influenced specific parts of the project more than others. The University of Connecticut houses the hardware and software, and the decision to use Endeavor Voyager and ImageServer to manage the database and images in phase I and use Endeavor ENCompass for this in phase II were results of their knowledge and experience. Mystic Seaport was experimenting with similar software and each learned through the development of these new technologies.

On another front, the Connecticut Historical Society's and Mystic Seaport's experience in museum education and relationships with primary and secondary teachers were crucial to the development of programs for teachers and students. Each developed summer institutes for teachers and encouraged input through the development of a teacher advisory board.

Expanding the partnership brought new ideas to the table and extended the scope of the partnership geographically. It was difficult at times for the representatives of the small institution, the New Haven Colony Historical Society, to attend meetings, as the small size of the staff and staff cutbacks during the project meant that often they could not leave their institution when meetings were scheduled.

The introduction of new, cutting-edge technology for this phase of the project also introduced some stresses, as there was difficulty in synchronizing the various processes involved in the addition of new images and metadata due to a greater than expected lag time in bringing the new technologies up to functionality. However, all those involved in the project carried on as efficiently as they could at every juncture, and throughout stressful periods, maintained focus on the ultimate goal and strengthened their determination to achieve it. As the end of phase II nears, the original partnership has solidified, building on success in phase I, and mellowed by virtue of the partners' enthusiasm and success.

1.2.3 Administration

With the awarding of the first grant, the partners established a Management Team with representatives from each institution plus the project director and project manager. The Management Team provided central direction to the project and served as a final arbiter for issues arising from other project teams. The project director chaired team meetings but there was no established hierarchy and representatives participated freely.

Early meetings highlighted differences both in individual personality as well as institutional cultures. University personnel tended to be more structured while museum and historical society personnel were more entrepreneurial. These differences complemented one another. The differences could have proven to be a stumbling block but the different groups quickly learned to value what each institution and individual brought to the project and this built an even stronger partnership.

1.2.4 Committee structure

As the project began, the Management Team appointed an Education Team to begin initial discussions with teachers to inform the selection process. This team met with an educational consultant and with individual teachers to gain input on what content and format would be most helpful and relevant to classroom teachers. Meetings later on in the project focused on the Web site itself. Through a strengthened teacher-project partnership, the organization and content of the Classroom section of the Web site was improved and enhanced (see section 2.2.3).

Soon afterward, a Selection Team began the work of defining the selection process and making the final selection of materials for digitization. They used input from teachers but were also guided by rights issues and criteria outlined in the IMLS grant proposal (see section 2.2.1). The Technical Committee also came together to coordinate digitization of materials, discuss metadata crosswalks and the implementation of Endeavor's ImageServer for the project.

The original committee structure has expanded in phase II of CHO to include Management, Selection, Technical, Cataloging, Education and Teacher Advisory Committees (Figure 1).

choII_org_chart

Figure 1. CHO organizational chart

2 Partnership with teachers

2.1 Identifying needs: teachers and students as target audience (how it began)

As part of the process of developing the grant proposal for the first phase of CHO, educators were enlisted to help define project goals. Bill Tally, director of the American Memory Project Fellows facilitated a meeting of project team members and a focus group of teachers from eight communities in March 1999. These educators were enthusiastic about the prospect of making primary source visual materials readily available for use in the classroom. They affirmed both the power of images to engage student interest, and the opportunity that working with them would afford for meaningful learning about the past. Access to these primary source materials would allow students to take on the role of historians, gaining understanding of the past through analysis and interpretation of its artifacts. The group suggested using thematic categories to select and organize the images so that connections between seemingly disparate objects could be discerned through appropriate lessons and activities and higher-order thinking fostered in students.

Once several hundred catalog records had been created in phase I of the project, a pilot Web site was developed to test the functionality of the database and the usefulness of the content for the target audience of Grade 7-12 teachers and students. A second focus group of teachers was recruited to review and critique the pilot site. They, too, were excited at the prospect of having their students take on the role of historians, but felt that there needed to be more contextual information and some narrative content to illustrate how these discrete objects revealed information about the past, and also examples of how CHO might be used in instruction. They hoped that CHO would provide activities to develop the relevant skills of historical analysis and synthesis, and also some models of final products that might result from using images as primary sources.

2.2 What it means to be user-centered, early outcomes

2.2.1 Selection

After the first grant was awarded, a committee composed of staff from the partner institutions used the input of the teacher focus group to define five thematic categories: Diversity, Environment, Infrastructure, Lifestyle and Livelihood. These were used as a lens for selecting Connecticut-related photographs, prints and drawings from the collections. They were also integrated into the descriptive metadata for the digital objects (Figures 2-6). As part of their process describing digital materials, catalogers entered the pertinent thematic category term or terms into the Subject.category element of the metadata record. This enabled searching and browsing of the metadata records in the search interface according to the five thematic categories. These themes also served as a framework in which to present educational content for the Web site, particularly the "Journeys" (see section 2.2.3).

mystic_CD2950_img0017.jpg

Figure 2. African American man in uniform; [189-?]

Taken by Everett A. Scholfield, a professional photographer who worked primarily in Mystic and New London, Connecticut from 1865-1913

Thematic categories: Diversity, Lifestyle

chs_CD0532_img0045.jpg

Figure 3. Interior of a blacksmith shop, Norfolk, Connecticut; [ca. 1890]

A man in striped jersey uses a hammer to shoe a white horse at the left. A second man, wearing a sweater, a jacket with frayed sleeves and a hat, holds the horse by the halter. A third man with a mustache and wearing an apron holds an iron bar on an anvil resting on a tree trunk. A fourth man, in a turtleneck jersey and plaid cap, holds a sledgehammer. The brick forge is in the right background. Tools and a large tub are nearby. A stovepipe connects to a chimney. Horseshoes are on the floor and hang from rafters. Other ironwork, including hinges, hangs from the wall at the left

Thematic categories: Infrastructure, Livelihood

mystic_CD4199_img0031.jpg

Figure 4. After half of ship hull, Groton Iron Works, Noank, Connecticut; 1917

Groton Iron Works had its main yard on the Thames in Groton and a second yard on the site of the former Palmer shipyard in Noank. It built a number of iron and wood-hulled vessels during and after World War I, including Emergency Fleet Corporation vessels. This corporation was formed ten days after the United States entered World War I in 1917. It was commissioned to build, equip, maintain and operate merchant vessels.The Groton Iron Works Noank yard ceased production in 1919. The Groton yard continued until 1928

Thematic categories: Infrastructure, Livelihood

uconn_CD1573_img0025.jpg

Figure 5. Women's basketball, Southern New England Telephone Company (SNET); 1928

Seven women posing with two men for team shot of the 1927-1928 Girls Basketball Team. Mabel Hubbell, "Doc" Hull, Edward Haesche, Catherine Eames, Hazel Hansen, "Buddy" Semack, Mary Porter, Eleanor Stribe and Julia Hull

Thematic categories: Infrastructure, Lifestyle

chs_CD0530_img0015_rev.jpg

Figure 6. Exterior of Gruenfeld's Market; [189-]

Three men wearing aprons and a woman stand in the doorway of a store. A sign reads: "Gruenfeld's / Meats / Fruits / Poultry". A sign stenciled on the windows reads: "Court Market". Meat and canned goods are displayed in the shop windows. An advertising sign reads: "Armour Star". A sign on the sidewalk reads: "Special / Legs of Lamb". Signs on the striped awning read: "Choice Meats" "Phila. Poultry". Two small girls with ribbons in their hair stand on the sidewalk

Thematic categories: Diversity, Environment, Livelihood

2.2.2 Cataloging

Substantial effort was invested in arriving at a common controlled vocabulary of subject headings as access points and a way of retrieving "more like this" results sets, including substantive "Browse" lists. It was decided to approach subject analysis and headings selection in terms of what would be most meaningful and accessible to teachers and students. In response to teacher feedback to the pilot Web site, greater attention was paid to creating descriptive titles and to using the "notes" area of the catalog record to provide contextual information for many images.

2.2.3 Educational content (Journeys, curriculum materials, classroom activities, citation guide)

Also in response to teacher feedback, classroom and photo essay sections were added to the plan for site content. Some members of the teacher focus group and project staff created lessons and activities that became part of the new Classroom section. "Journeys", or photo essays suggested by the original teacher focus group, were created by project staff and interns. These were organized in the thematic categories and provided a historical context for selected images and suggested ways in which the thousands of item-level records can be used to reconstruct the past and tell a story. Guidelines for Citing Resources in Connecticut History Online was added to the Classroom area of the Web site as well, offering examples of citing images, Web sites and texts, and links to online reference sources.

2.2.4 Web site and search interface design

The official project launch in January 2002 attracted a number of teachers who affirmed the value of CHO as an educational resource. Teacher workshops at two of the partner institutions organized in the summer of 2002 also provided positive feedback about the site and ideas for further development of the educational component in future phases.

2.3 Maintaining, evolving and deepening that partnership

2.3.1 Partnering outcomes in phase II

Based on comments from teachers in phase I, the partners made it a goal in phase II to establish a long term, collaborative relationship with teachers to improve the site and make it more relevant and useful as a curriculum and classroom tool. The first step was the formation of two key committees: a Teacher Advisory Committee that would identify curriculum goals, provide feedback on educational resources mounted in phase I, and aid in the development of the Web interface and additional Web content; and an Education Committee comprised of education staff from all five partner institutions.

Members of the CHO Education Committee, including curators responsible for instruction and museum educators, took on the tasks of raising awareness about the project among teachers, curriculum coordinators, and school media specialists in the state. The initiatives pursued by the committee include: collaborating with the Teacher Advisory Committee; exploring and reporting educational standards to project staff; drafting editorial guidelines for educational content on the Web site; and suggesting ideas for developing and sustaining CHO into the future.

A particularly fruitful effort on behalf of the Education Committee was hosting an open house for teachers. The event took place in the Connecticut Historical Society Museum located in the center of Connecticut and attracted 25 attendees, among them middle and high school teachers, elementary school teachers, school library media specialists, instructional technology specialists, and two professors from University of Connecticut's Neag School of Education. The program included a CHO demonstration, a report on forthcoming features and new materials, a lively question and answer session, hands-on laptop stations, refreshments and plenty of personal dialog. In addition to raising interest in and understanding of the project, the event resulted in an increase in membership and an expansion in the variety of experience represented on the Teacher Advisory Committee, as well as a host of suggestions for enhancing the usefulness and usability of the Web site.

2.3.2 Joint projects

In addition to the open house, a direct-mail brochure sent to social studies teachers around the state, and the integration of CHO into educational programs at partner institutions, generated further opportunities for collaboration with teachers and schools. Since the event, CHO has been invited to present workshops at several teacher conferences, submit a grant proposal to fund a teachers institute, and present at regional professional development workshops for teachers. CHO was also recently selected for inclusion in the National Endowment for the Humanities' EDSITEment project, a Web site that brings together educational and humanities resources from cultural institutions worldwide.

CHO in phase II is currently based at the University of Connecticut's Dodd Research Center. Establishing a campus presence, together with the Education Committee efforts to recruit university professors to participate on the Teacher Advisory Committee, have improved communication and opened up partnership opportunities with several academic departments. Forthcoming projects include a School of Education internship program for Master's candidates. Student interns will work with a local school district, CHO project staff and curators at the partner institutions to research and develop programs and materials that foster effective use of primary sources in the classroom. New educational content created in the course of the program would become part of the site.

Another project, part of a major award funded by the Carnegie Foundation, involves a collaboration between the School of Education and College of Arts and Sciences. As part of their preparation as subject specialists, graduate students will work with educators at a local high school to provide training in the use of CHO and develop and test additional curriculum materials that incorporate the site.

3 Coming to common terms: collaborating on metadata

3.1 Function/goals of metadata standards broadly and in CHO

Reaching consensus among five institutions on best practices and metadata standards used to describe the digitized materials was a significant part of the effort to create a centralized resource and provide coherent access to materials. Decisions on cataloging practice stemmed from partners' commitment to a primary mission. From selecting materials to describing and delivering them, the purpose of the project was to develop a resource that would enhance access to and use of cultural heritage materials. The catalogers therefore were charged with describing materials in such a way as to reveal how the people of Connecticut lived, worked, traveled, and engaged in national and international events through time.

Defining a target audience and educational objectives was necessary to establish common practice. The user community was identified as students, teachers and the general public. Partners agreed that the Web site would serve as a starting point for broader research, and enable the exploration of history in an integrated, seamless environment. Teachers would be provided with online resources--database, lesson plans, classroom activities, and historical essays, illustrated with images and documents from the partner institutions--to identify and better integrate primary source materials into their curriculum. Materials would be described and contextualized in such a way that would encourage students to search for and make connections between their local communities and the universe of primary source materials that document them. It was agreed by the partner institutions that the site needed to be relevant and engaging, support meaningful inquiry, and enhance the general public's familiarity with archival and museum collections at partner institutions for CHO to be sustainable and fulfill its mission to enhance access and use.

Dublin Core was selected as the appropriate element set for implementing and maintaining metadata in this project. Its simplicity and extensibility suited our need to assemble and integrate records from different systems into a centralized system. As an international standard that can be stored and delivered as XML, it was viewed as a sustainable option, one that would enable us to share metadata, more readily build (for instance as an OAI data provider) or enhance existing digital library systems, and benefit from development tools, both locally and worldwide.

3.2 Building on a common vision of and commitment to access: CHO cataloging committee and creating best practices

In addition to enabling and improving resource retrieval and discovery, cataloging guidelines were developed to provide quality control of metadata and ensure interoperability with other descriptive standards being used in the museum, library, and archival fields. Guidelines and best practices were crafted in ongoing discussions among Cataloging Committee members and mounted on the CHO Web site along with other project cataloger resources such as checklists, tips on using the client application, links to thesauri, meeting minutes, and other documentation.

The emphasis in cataloging practice has been on applying multiple access points and terminology that is as accessible as possible to the primary target audience, middle- and high-school students and teachers (Eakins and Graham 1999). Within this conceptual framework, catalogers utilize appropriate existing standards and tools, such as established subject thesauri, to the extent possible (Digital Library Forum 2001). In maintaining a focus on our audience, catalogers choose more intuitive terms over more arcane ones, and apply multiple headings to express facets of an object (trying to minimize the use of subdivisions) that will enable searching and browsing by broad topic and genre. Multiple notes fields are used to provide as much explanatory and contextual information as possible.

Guidelines were often modified to reflect regular cataloging discussions and, sometimes, decisions that had to be negotiated among all members of the Cataloging Committee. As part of their process, many catalogers had to modify MARC-based or museum system records for inclusion in CHO. Ultimately, catalogers worked together as a team, acknowledged and shared expertise from their differing professional backgrounds (academic library, museum, public library), and formed consensus on practice in order to meet the needs of the project audience.

3.3 Impact of software decisions

In the first phase of CHO, Endeavor's Voyager system was used for storing, searching and delivering metadata, with Endeavor's ImageServer as the mechanism for storing and accessing their associated digital images. ImageServer made it necessary to create MARC records in the Endeavor Voyager system rather than Dublin Core records, although most cataloging was done using a Dublin Core-like template. For display, Dublin Core elements were therefore mapped to MARC, using a crosswalk.

In phase II, metadata was mapped to qualified Dublin Core and migrated to ENCompass for Digital Collections. Today catalogers create new records for a variety of materials directly in the system using a desktop client application and a qualified Dublin Core template that has been modified slightly for the purposes of the project.

Technical and workflow issues presented themselves during the first four months of implementation. In addition to filling project goals, project partners needed to adapt local cataloging workflows and practices in order to accommodate the relatively experimental aspects of centralizing descriptive metadata, and the lack of tools provided by the digital library system (ENCompass) for easy data management and manipulation. As more records were added to the database, catalogers refined their practice and often revised records accordingly, with earlier decisions being superseded by more recent ones.

3.4 Staffing and supervision

During phase I of the CHO project, funding from IMLS allowed each of the three partners to hire their own cataloger/metadata specialist. The management team discussed hiring staff centrally but ultimately rejected this arrangement due to the geographical distance between institutions and the differing administrative structures. Each institution had significantly different hiring practices and benefit structures that could not be bridged. Initial hiring was coordinated between the institutions while taking into account human resource requirements.

During phase II, four of the five partners received support to hire catalogers while one institution integrated the cataloging responsibilities into its normal institutional workflow. Supervision remained with the individual institution. The partners updated and revised the cataloging standards, but the effort of the initial phase of the project meant that there was significantly less time and effort required to direct the work of the catalogers.

During phase II, one of the partners had difficulty hiring a cataloger. As an alternative, the Management Committee came up with the solution of extending the tenure of the experienced cataloger at one of the other institutions to catalog those materials. This solution entailed transferring funds from one partner to another but will result in the outcome promised in the IMLS grant.

Once hired, each cataloger received an orientation from the parent institution and was supervised by an institution staff member. The project director and project coordinator provided program oversight and cataloging benchmarks, training in use of the cataloging client application, proofreading of catalog records, and ensured that catalogers participated in regular catalog meetings and project listservs.

4 Technology: agreeing on digital formats, delivery tools and systems

4.1 Models for delivery

4.1.1 Endeavor Voyager and ImageServer in phase I

The project used Endeavor Voyager, an integrated library system (ILS), to manage the database and MARC metadata records. Endeavor Voyager was a new product when it was selected for phase I but has since become an industry standard for ILS at most major research libraries throughout the United States. CHO operated a separate database on a University of Connecticut server.

The project used Endeavor ImageServer software to capture metadata for digital images and display those images. Dublin Core was originally selected as the cataloging standard for phase I, although ImageServer was not capable at that time of storing or exporting descriptive data in the Dublin Core format. As a result, it was decided to map the Dublin Core fields to MARC fields using a template developed within the ImageServer system.

4.1.2 Endeavor ENCompass in phase II

Between phases I and II of CHO, Endeavor Systems developed a new product called ENCompass. Included in this suite of products was a module called ENCompass for Digital Collections. This product allowed users to create digital collections using a variety of media and metadata standards. Both the University of Connecticut and Mystic Seaport were interested in using this new product. The University of Connecticut purchased this module and the partners agreed to use this for metadata creation and delivery of digital images.

4.2 Allocating costs for and carrying out digitization

4.2.1 Digitization in phase I

In phase I, partners outsourced the digitization work to a corporate vendor, Boston Photo Imaging (BPI), at a cost of $3.50 per image. This solution met all partner criteria since the cost was reasonable, partners could maintain their collections on site and avoided the duplication of equipment and expertise. In addition, direct scanning of the original artifacts did not lend itself to the capture of a large number of images in a short period of time.

BPI spent up to one week at each institution to create film negatives of each standard size project image (up to 11 x 14 inches). The inter-negatives were then converted to digital format in the form of Photo-CDs. For a small number of images larger than 11 x 14, materials from all institutions were assembled at Mystic Seaport for copying. This process facilitated the capturing of large quantities of images without having to stop and make the constant adjustments that direct scanning would require. Each image was recorded on the photo-CDs at four resolution levels. For the purposes of public access, and to ensure image security, only two levels of image resolution are available on CHO: a low-level thumbnail image, and a second mid-range resolution image.

Converting the images to digital format required the following steps:

BPI's visits were scheduled towards the beginning of the project. Since only a portion of the graphics at each institution had actually been fully described by then, an accession or control number was assigned to every graphic so that photography could proceed.

4.2.2 Digitization in phase II

In phase II, partners agreed to again outsource digitization to BPI. Following our initial selection, we discovered that oversize images done by BPI would limit the amount of material that could be included in the project. We agreed to use BPI for all standard size images up to 8 1/2 x 14 but to develop other alternatives for oversized material. The Connecticut Historical Society had developed a good working relationship with a Hartford-based vendor and they digitized 655 images using that vendor.

Since CHO phase I, Mystic Seaport has moved to a new curatorial building that includes a full-service digital lab. Mystic Seaport indicated that they could digitize oversize images for other project partners at a reasonable cost, and they will carry out work on oversized material from their own collections, and those of the Connecticut Historical Society Museum and the University of Connecticut.

4.3 Developing new technology

4.3.1 Geolocator

The Geolocator is a unique feature of CHO. In an attempt to make searching a bibliographic database more appealing to a wide audience, project staff created a map of Connecticut with dynamic links between images and their records. This required evaluating the efficacy and structure of the geographic fields in a MARC record for use with a geographic information system. Using an ambiguous search (searching for all instances of a name), the Geographic Name field (651) was inefficient. By pre-processing all the names in the 651 field, a more specific search was achieved.

Tables were created after an initial polling of the geographic names in the database. URLs are constructed from the place names and then geo-referenced against appropriate thesauri using ArcView. To take into account historical names and places that no longer exist, the Connecticut Digital Gazetteer was created as one of the GeoLocator's thesauri.

The result of searches produces a Connecticut map that includes graphical representations of the geographic name fields in the image records. Researchers are able to view this map, highlight any place or group of places, drilling down as far as street names and numbers, and then viewing the resulting set of records and images. In CHO phase II, searches in the GeoLocator system will be further refined and some of the problems with historical place names will be resolved.

4.3.2 JPEG2000

A major goal of the project in phase II was to explore and incorporate new technologies to enhance the delivery and use of digital materials in the CHO collection. Early in phase II, the Technology Committee discussed possible methods for storing and delivering large format visual materials such as maps, charts, newsprint, and broadsides. Two partners had experience applying the .SID format for this purpose.

At the time, Rutherford Witthus, Curator of Literary and Natural History Collections at the Dodd Research Center, University of Connecticut, began using JPEG2000 for Charles Olson's Melville Project, a collection of digitized handwritten manuscript materials from the Charles Olson Research Collection. JPEG2000, a new, open standard (non-proprietary) technology, allows for lossless compression of large digital files and is a means to embed XML documents/metadata in the file itself. When viewed in a JP2 viewer application, users are provided with tools to zoom and pan in the digital object.

The UConn Library had successfully compressed digital image files, and embedded an Encoded Archival Description (EAD) finding aid, a PDF document, and other descriptive data in the file, using JPEG2000. The Technical Committee agreed that although JPEG2000 was a new technology that had few models for development and use in the library field, its benefits and status as a non-proprietary standard made it a better, more sustainable solution than SID technology for delivering and viewing large format digital materials. As a result, with UConn's Library's support and tools developed by Aware Inc., the project decided to implement JPEG2000 and share expertise of this new technology among partner institutions.

5 Forward momentum: opportunities for future collaboration

5.1 Developing project sustainability

Project sustainability is a major issue with digital projects developed over the past decade. IMLS and other funding agencies encourage projects to consider this issue and to develop strategies to address this need.

During the second phase of CHO, the Management Team began reviewing different models for sustainability. Fortunately for CHO, others are beginning to address this issue and there is help available from the Council on Library and Information Resources and others (Bishoff and Allen 2004, Zorich 2003).

The Management Team investigated different organizational structures and reviewed future income required to maintain and develop CHO over time. It is currently developing a long-term business plan and will probably establish CHO as a separate non-profit corporation in the near future.

5.2 Establishing a central rights management program

One difficulty with a multi-institutional project is fulfilling customer orders that involve collections from different institutions and managing rights for copyright and use. The current solution has been to refer individuals to the institution holding the original material and establishing a relationship between that institution and the customer. This solution is not customer-friendly when materials being sought come from different institutions with differing costs and procedures.

CHO has proposed developing a centralized rights management and fulfillment system that would coordinate efforts among the institutions and provide customers with "one-stop shopping" for both their reproduction requests and rights approval. This proposal was developed for IMLS as part of the library leadership grant proposals for 2004. If this is funded, it will become part of a larger strategy to develop financial support for CHO and further the project's effort toward sustainability.

Developing such an effort will push the partnership towards increasing levels of cooperation and require even greater communication and compromise to achieve common approaches toward fulfillment and rights management. Such discussions will test the strength of the partnership. However, if this is successful, it provides CHO with another "value added" service that will encourage other institutions to join the CHO program. It also provides a model that can be shared with others in the museum, library and archives communities.

5.3 Building project-teacher momentum

Developing good communication with the educational community requires an ongoing effort. Information gathered during phase I of CHO was critical to designing a site that met teacher needs. Initial teacher meetings during phase II identified teacher needs and suggested improvements that will appear on the CHO site during its second iteration.

Teacher meetings will continue after the new site appears, and the project is planning programs that will allow for easy additions and updates of curricular material. We also look forward to engaging in a new relationship with the NEAG School of Education, which recently received a $5 million grant from the Carnegie Foundation to help improve teacher education, and to participating in other programs currently in development with teachers around the state.

6 Conclusion

As the largest database of historic images of Connecticut combined with materials for teachers and students, CHO has already had an enormous impact on Connecticut's historical and educational community. Its success can be attributed to the skills and experience of the staff participating in the project and the teachers and educators who have provided input on the project and process. However, without the emphasis on partnership - partnership between the grant participants and partnership between teachers and the CHO project members - much of that energy might well have been lost or misdirected. As the project moves toward a new phase that requires a new vision of a sustainable effort, partnership is a value that continues to be critical and will allow the project to move to a new level of success.

Acknowledgements

The authors wish to thank the Institute for Museum and Library Services for their grant support of Connecticut History Online over two separate granting cycles. The authors also thank their parent institutions, the Connecticut Historical Society Museum, the Connecticut State Library, Mystic Seaport, the New Haven Colony Historical Society and the Thomas J. Dodd Center at the University of Connecticut, as well as their staff who served on project committees or carried out work so that others could participate in Connecticut History Online.

References

Bishoff, L. and Allen, N. (2004) Business Planning for Cultural Heritage Institutions, CLIR Publications, No. 124, January (Washington, DC: CLIR) http://www.clir.org/pubs/abstract/pub124abst.html

Digital Library Forum (2001) "A Framework of Guidance for Building Good Digital Collections", IMLS, November 6 http://www.imls.gov/pubs/forumframework.htm

Eakins, J.P, and M. E. Graham (1999) "Content-based Image Retrieval: A report to the JISC Technology Applications Programme", January http://www.unn.ac.uk/iidr/research/cbir/report.html

National Council for the Social Studies (1994) Expectations of Excellence: Curriculum Standards for Social Studies, January (Silver Spring, MD: National Council for the Social Studies) http://www.ncss.org/standards/stitle.html

Riley, C. (1999), "Evidential Understanding, Period, Knowledge and the Development of Literacy: A Practical Approach to 'Layers of Inference' for Key Stage 3". Teaching History, Vol. 97, November, 6-12

Wineburg, S. (2001) Historical Thinking and Other Unnatural Acts: Charting the Future of Teaching the Past (Philadelphia: Temple University Press)

Zorich, D. M. (2003) A Survey of Digital Cultural Heritage Initiatives and Their Sustainability Concerns, CLIR Publications, No. 118, June (Washington, DC: CLIR) http://www.clir.org/pubs/abstract/pub118abst.html