Genesis
Within the context of Internet-wide interest in access to distributed resources, Gallery Systems’ CEO Jay Hoffman designed a tool in 2000, that would allow multiple collections to be searched simultaneously over the Internet from a single entry point. This approach is called a “distributed search.” A search is described as “distributed” because the data may be stored in multiple databases, each of which is being searched in parallel. These databases may be physically located at different sites (for example, on servers at individual museums), as long as they are connected via the Internet.
Figure 1: eMuseum Network logo.
The First Pilot Project
Our clients received the concept of distributed search as an exciting opportunity to make key collections information available to one another for research and collaborative projects. While much of the information may already be available on individual museum Web sites, the ability to search and retrieve information from many museums in a single search and to view it in a consistent format had been a goal in the museum information community for some time.
As part of Phase I of its Distributed Search Pilot Project, Gallery Systems and 10 participating museums successfully tested the technical and practical application of cross-collection searching of metadata as a means for museums to share collections information with each other. It became clear that Distributed Searching held promise for museum professionals in a range of activities, including information sharing, collaboration and research.
Figure 2: Partners of
the first Distributed Search project
Moving Forward a Network
With the launch of Phase II of the project, the eMuseum Network connects directly to an easily updatable snapshot of an institution's database hosted locally at the institution. A participating institution does not need to be a Gallery Systems client, a TMS or an eMuseum user. Based on an improved Distributed Search technology, the eMuseum Network has minimal impact on time and resources needed by the participants.
The new Web site, http://www.emuseum.net, provides a single access point for searching a selection of collections from participating institutions. Search screens allow users to select specific fields to search on (including Keywords) and specific collections to include or exclude in the search. Search results are returned in a unified group, with all objects presented in a common, unified format. Examples of possible searches include queries for a specific artist (“Cezanne”), queries based on medium or time period (e.g., French pastels), or queries based on descriptive or content fields (“Annunciation”).
Figure 3: eMuseum Network login
page.
Who Can Participate?
Potentially, any cultural institution can participate, but participation in the eMuseum Network is currently by invitation only or upon request. Only institutions that are sharing collections data will be allowed to enter the network to access data.
Each participating institution can decide how much of their collection to make available. For example, it would be possible to include data only on objects that are already included on a public Web site or to make more data available to a scholarly audience.
Participants will be asked to make collections information available and we plan to include a forum for discussion so that participants can share ideas.
The goals of the second phase of the project build upon Pilot I:
- Improve ease of installation, especially related to data mapping
- Test performance through increase of participating “nodes” to 25 (all hosted locally by institutions)
- Explore interface enhancements to ensure user satisfaction
We are now in the process of inviting several of our clients as well as other institutions to participate in the eMuseum Network Project Phase II, and scheduling the installations of our former pilot project participants.
Participating museums will have an opportunity to test this technology first-hand and to provide feedback to improve its usefulness for the future. Resource requirements for participation are minimal, since a Java installer is provided and maintained by Gallery Systems’ product management team.
The installation of “nodes” (software components on the client Web server) has begun in March 2007. If the institution has limited human resource available at the time of installation, Gallery Systems can use Remote Access to allow our technicians to quickly implement the program components. As the project comes online, participating institutions will be invited to explore the tool and provide feedback through surveys and other means.
Current And Potential Participants
The current participants are:
- George Eastman House International Museum of Photography and Film, New York
- International Center of Photography, New York
- National Portrait Gallery, Washington D.C.
- The Frick Collection, New York
- The Metropolitan Museum of Art, New York
Architecture Overview
Basic Topology
The search engine accesses secure databases over the Internet.
The eMuseum Network is based on software components developed by Gallery Systems using World Wide Web Consortium (W3C) and other standards. These include Web Services, Simple Object Access Protocol (SOAP), and Extensible Markup Language (XML), the Java language, Java Database Connectivity (JDBC) and open databases.
The architecture includes a central server that provides the user interface through a Web browser and maintains a registry of databases available for searching. The central server provides the secure connection to the individual remote databases (“nodes”) and is transparent to the end user performing the search.
Figure 4: eMuseum Network Topology.
Workflow
- A user visits eMuseum.net Web site and enters a query
- he query is transmitted via HTML form to GS eMuseum Network Server running ASP/COM through IIS.
- The COM component on eMuseum Network Server packages this query into a SOAP request and transmits it to Java Web Service on GS Central Server running Apache Tomcat.
- The Central Server forwards the SOAP request to the Java Web Service on “Node Server” at each participating institution (the Node Servers are also running Apache Tomcat).
- The Node Server translates the request into a SQL query and queries the local database via JDBC.
- The Node Server packages JDBC result set into a SOAP response and transmits back to the Central Server.
- The Central Server aggregates results from different nodes into a single SOAP response and transmits it to the eMuseum Network Server.
The eMuseum Network Server extracts XML data from the SOAP response and uses XSL style sheets to generate HTML for the user.
More about the User Experience
The search interface can be launched from the following URL:
In a simplified example, a search is initiated by a user as a single query, which is sent by the search engine to each of the databases specified in the interface (participating member databases). Each of these databases then executes the search and sends the results back to the search engine. The search engine then combines the results from each of the databases into a single result set and displays it to the user.
The following user scenarios are based on eMuseum Network.
User Scenario – Advanced Search
A scholar is conducting a survey about early photographs and direct image making processes (with no negative originals). For instance, searching for daguerreotypes, a technical process named after the artist (and chemist) Louis Daguerre (1787-1851).
With eMuseum Network, he could use the Advanced Searchoption and access many resources conserved and curated by the International Center of Photography and the George Eastman House, in particular portraits, comparing long exposure times, lenses, heat of the cup of mercury that was used by Daguerre, densities within the final photograph, etc.
Figure 5: Advanced Search.
Figure 6: View 8 results at a time.
Figure 7: View a list of results withthumbnails.
Figure 8: View detailed catalog.
Figure
9: View Image.Portrait of a blind man holding a cat - © George Eastman
House International Museum of Photography and Film
User Scenario – Simple Search
User is studying the Sino Japanese war and wants to access resources related to a photographer, Robert Capa, in particular his Sino Japanese war photographs in between 1932 and 1938, from two participating museums, the International Center of Photography, and George Eastman House International Museum of Photography & Film.
Figure 10: Simple Search
User Scenario – Keyword Search
User can search artworks using simple keyword search in the box that appears on the upper right side on any page of the user interface, as shown on Figure 11 below.
Figure 11: Keyword Search (flowers) and Label Copy
View The Object On Original Web Sites
When it’s possible, the catalog view within eMuseum Network contains a link to directly access the object described on the source Web site:
Figure
12: Venus of the Frick Collection – link to
original Web site
Figure 13: Venus of the Frick Collection on the original Web site http://collections.frick.org
with another Gallery Systems product (WebKiosk)
The Central Server
The Central Server is the heart of the eMuseum Network, which encapsulates the business layer and is responsible for:
- Receiving SOAP requests for the Web server, transmitted as Java Web service
- Forwarding SOAP request to Java Web service on a “Node Server” at each participating institution
- Aggregating results from different nodes into a single SOAP response and transmits it to the Web Client
The Central Server is Web application developed in Java/XML, running Tomcat as application server.
Distributed Search Node
The Distributed Search Node acts as a gateway to a remote collection database. The Central Server generates an abstract query and distributes this to the Node Server. On the Node Server the abstract query is transformed into a database specific query format (SQL) and run against the selected database instance. The result set is encoded into an XML document.
Figure
14: Software components and workflow
About Photomuse
Figure 15:Photomuse http://www.photomuse.org
The eMuseum Network is not the only application of Gallery Systems’ Distributed Search technology. The International Center of Photography (New York, NY) and George Eastman House (Rochester, NY) have launched a collaborative Web site using distributed search technology, a resource for scholarship in the history of photography. ICP and Eastman House (the world's oldest photography museum) are collaborating on a portal to allow Web access to the collections of both organizations, with a plan to include other photographic resources in the future Earle and Bruce (2004). The two institutions have successfully tested the same query system that powers eMuseum Network. The technology allows Web visitors to seamlessly search multiple databases from a single search screen and discover “Photography at Work”. The prototype of this recreational scholarship resource, called Photomuse, was rated as one of the 50 coolest Web sites 2006 by Time Magazine (July 26, 2006).
Conclusion
The eMuseum Network:
- provides a mechanism to specify the depth of data that can be searched and displayed (it is not an “all or nothing” rule). For example, a museum can allow a search on object names, but display only titles and description in the set of results.
- provides access to the most current data extracted from a participating institution’s collection management system.
- is a dynamic tool – enhancements could be developed as part of the main platform, but also as separate open source initiatives through an application programming interface, as dictated by individual needs in our community. Among these enhancements, we can envision common virtual exhibitions, loan requests, joint acquisitions, access to the network from or within collection management systems, advanced educational and research tools, accessible from anywhere.
- empowers search for objects, people, media, and exhibitions.
- allows the emergence or consolidation of spontaneous communities, based on similar collections or interests and/or common geographical or political areas.
- allows these communities to launch projects and recruit new participants through the eMuseum Network.
- enables comparisons between different collection management systems, specifically Gallery Systems’ TMS and future Web-based cataloguing products, software from other vendors, or open source systems.
- enables comparisons between different techniques of cataloging and several levels of subject descriptions (in the area of photography for instance).
- provides a central method to access acute information.
Acknowledgements
The author would like to thank Edward Earle, Curator, Collections / Director of Information Systems, International Center of Photography, for his invaluable support and dedicated energy to the Distributed Search technology and eMuseum Network.
The eMuseum Network could not exist without museum community members who advocated the idea of sharing collections online: Roger Bruce, George Eastman House International Museum of Photography and Film, New York; Edward Earle, International Center of Photography, New York; Linda Thrift, National Portrait Gallery, Washington D.C.; Vivian Gill, The Frick Collection, New York; Doralynn Pines, The Metropolitan Museum of Art, New York.
In addition, amongst the many institutions with whom we are currently exploring solutions, we would like to especially acknowledge the following: McCord Museum, Montreal, Canada (Nicole Vallières); Dallas Museum of Art, Dallas, TX (Jessica Heimberg); The Saint Louis Art Museum, Saint Louis, MO (Ella Rothgangel); Museum of Fine Arts, Boston (Linda Pulliam); Harvard University Art Museums, Cambridge, MA (Gillian McMullen); San Francisco Museum of Modern Art, San Francisco, CA (Marla Misunas); Archives & Museum Informatics, ICHIM and its committee of experts in cultural computer science.
References
Earle and Bruce (2004). Earle, E. W. and R. Bruce. Pictures and People: Distributed Query Database Collaboration. In D. Bearman and J. Trant (eds.). Museums and the Web 2004: Proceedings. Toronto: Archives & Museum Informatics, 2004. last updated March 25, 2004, consulted July 26, 2007. http://www.archimuse.com/mw2004/papers/earle/earle.html
Maryanne Murray Buechner, Time online in partnership with CNN, section BUSINESS & TECH, 50 COOLEST WEBSITES 2006, Entertainment, Arts and Media, last updated July 26, 2006, consulted July 26, 2007. http://www.time.com/time/business/article/0,8599,1219395,00.html
Cite as:
Rebuffet, J-P., The eMuseum Network: Searching Shared Collections , in International Cultural Heritage Informatics Meeting (ICHIM07): Proceedings, J. Trant and D. Bearman (eds). Toronto: Archives & Museum Informatics. 2007. Published October 24, 2007 at http://www.archimuse.com/ichim07/papers/rebuffet/rebuffet.html
Search A&MI