Access in the Distributed Environment: Making On-line Museum Content More Accessible to Educators
The Project Team
The Canadian Heritage Information Network (CHIN) identified resources, defined an action plan, and managed the development of this gateway application. CHIN was responsible for defining the architecture/structure of the site and application. As the host of the gateway, CHIN will ensure ongoing maintenance, promote continued growth and continued improvement of the application. In the initial stages of the project, CHIN formed a project team by contacting museums and others who had expressed an interest in or had certain expertise in museum learning materials, the Canadian curriculum, or distributed access tools.
The project team also included five interns. CHIN manages a component of the Department of Canadian Heritageís Young Canada Works Internship Program, part of Canadaís Youth Employment Strategy. Through CHIN, internship funding was provided to those organizations participating in the project that were interested in hiring interns and demonstrated the ability to provide valuable work experience and the necessary supervision. See Annex A for a list of project team members and their contributions.
Project Goal and Objectives
The goal of the Learning with Museums project is to make on-line educational museum content more accessible to educators, students and life-long learners. In the Learning with Museums project, museums create a metadata record for each of their on-line resources, using a cataloguing tool provided by CHIN. CHIN sends the metadata back to the museum, and they embed it within their resource. CHIN can then harvest this embedded metadata periodically in order to update its metadata repository. Although the resources in Learning with Museums are distributed across various servers and sites, the use of metadata makes it possible for users to perform powerful searches of the learning materials as if they were part of a centralized resource. The metadata embedded in the educational resources also has the potential to enhance Internet search accuracy, improve resource discovery capabilities, and facilitate cross-domain interoperability.
Users of Learning with Museums are then able to directly search the metadata in order to quickly and easily find the educational resources that meet their needs. When an appropriate educational resource is found, users will link to it directly.
Several objectives for the project were defined:
Review of Existing Models
As a point of departure, existing models for accessing distributed on-line learning materials were investigated. The first of the models is the Learning Resources section of Canada's SchoolNet (http://www.schoolnet.ca), which offers access to learning resources by curriculum areas, grade levels, themes, and keywords. As SchoolNet deals with Canadian curriculum and provides bilingual access, CHIN's project team found it to be a useful model.
The Gateway to Educational Materials (http://www.thegateway.org) was also examined as a model for Learning with Museums. This initiative of the National Library of Education and the U.S. Department of Education was created in 1996 "to create an operational framework to provide America's teachers with 'one-stop, any-stop' access to Internet-based educational resources." As the Gateway to Educational Materials (GEM) is in its third year of development, CHIN's project team was able to observe this very impressive gateway, as well as its related tools and user guides, in operation.
CHIN also studied several other projects which enable access to educational materials - SCRAN (Scottish Cultural Resources Access Network, http://www.scran.ac.uk/), EdNA (Education Network Australia, http://www.edna.edu.au/), and ADAM (the Art, Design, Architecture & Media Information Gateway, http://www.adam.ac.uk/). Included in this review was an evaluation of their guidelines and tools to catalog/tag resources, search methodologies, interfaces, retrieval strategies and access points.
As a result of the analysis, aspects of these models were incorporated into the development of the CHIN application.
Part of the investigation of the existing models was to determine whether there was a niche for a gateway that contained on-line museum content exclusively. Due to the authoritative nature and high quality of the content produced by museums, it was determined that providing access to this material would be a benefit to all Internet users. The on-line educational content developed by museums is often complemented by their in-house educational programming, activities, and events; this broadens the potential use and value of museum on-line material.
Metadata plays a very important role in the Learning with Museums project. CHIN chose to adopt a metadata schema based on that of the GEM project (which is in turn based on the Dublin Core 15-element data). CHIN's work was also informed by the metadata schema used by Canada's SchoolNet.
CHIN has been involved with developing guidelines for using the Dublin Core in the museum and heritage context through its participation in the Consortium for the Computer Interchange of Museum Information (CIMI) testbed projects, and is currently using Dublin Core metadata in several of its resources. CHIN wanted to continue to conform as closely as possible to this international standard.
As the Learning with Museums project started, the Dublin Core Metadata Initiative was in the process of defining a set of "qualifiers" to the 15 core elements. These qualifiers were to refine the meaning of the elements in order to allow more precision. For the Learning with Museums project, user requirements were such that it was necessary to use Qualified Dublin Core to enable precision in retrieval. Although the Dublin Core Qualifiers were not yet finalized by the time the project started, CHIN adopted a set of qualifiers based on current Dublin Core Qualifiers working drafts.
In addition to the Dublin Core Elements and Qualifiers, CHIN adopted metadata elements from the GEM schema, which contains eight education-specific elements along with the Dublin Core 15-element data set (see Annex B).
A crosswalk between the SchoolNet, GEM metadata, and CHIN metadata schemas was created, and the CHIN metadata schema was shared with Canadian Studies, which is currently developing a database of off-line educational materials. This will enable interoperability at various levels between Learning with Museums and CHIN's project partners.
The project team faced several challenges when trying to find a way to express the metadata so that it could be interpreted by harvesters. The Learning with Museums project required that complex metadata (see Annex B) be embedded in HTML documents while being compatible with all versions of Web browsers. Use of html metatags would not enable the expression of a complex data structure in a standard way; there was a lack of syntactic rules, creating ambiguities which the parser could not overcome. There was no syntax to express groupings of elements, given that there were repeating elements and qualifiers a syntactic qualifier was needed. Without a standard, no harvester would be able to interpret the metatags in a reliable way.
The chosen solution was to embed XML (Extensible Markup Language) in HTML documents using RDF (Resource Description Framework). With XML, CHIN was able to utilize the CIMI Dublin Core DTD (Document Type Definition) and create a CHIN DTD that defines a more complex data structure. RDF provides the syntax for DC, GEM, and CHIN elements; and the DTD provides the grammar and the structure that will allow other resource discovery tools to interpret the RDF. The CHIN project team used "abbreviated RDF", so that the RDF metatags could be embedded in the museums' html Web pages; if "serialized RDF" had been used, extraneous RDF information outside of the RDF tags would be displayed when using some of the older web browsers.
An important consideration in the choice of RDF was an article entitled, "Guidance on expressing DC in RDF". This article is a Working Draft of the Dublin Core Metadata Initiative (DCMI) which has been offered to the Dublin Core Advisory Committee for acceptance and ratification. This document states that "...only the combination of RDF and XML currently offers the range of capabilities required to fully express the richness of both qualified and unqualified Dublin Core in an unambiguous manner." It was important that CHIN follow the direction of the DCMI to allow for future interoperability.
Another consideration was that XML and RDF are World Wide Web Consortium (W3C) standards, which means that there is a large user community. Other users have created tools using these standards, which can be adopted or used as models. Another benefit of using a W3C specification is that there is ample resource material available on-line, which is very helpful during the development process.
Dublin Core, GEM, and SchoolNet have each created controlled vocabularies and/or format standards to be used with their metadata element sets. CHIN was able to adopt or adapt many of these standards. However, as CHIN is dealing with bilingual content, some of the existing vocabularies had to be translated in order to enable bilingual access to the resources. A thesaurus was also created to deal with bilingual subject access.
In order to create a subject browser to be used in both cataloguing and retrieval of the learning resources, subject terms were drawn from each provincial and territorial curriculum. Due to the fact that some provincial/territorial curricula are currently under review, the application will need to be updated as new protocols are defined. As well, for the first phase of the project, only subject terms drawn from grades 7 -13 curricula were included. As this is a dynamic and evolving application, in subsequent phases, the project will also look at curricula below grade 7.
Although controlled vocabularies are used extensively in this project to ensure resource discovery, the resources can also be described with uncontrolled vocabulary to allow freedom of expression and alternatives to the controlled vocabulary.
The Cataloguing Tool
In order to assist museums with the task of embedding metadata in their on-line educational resources, CHIN created a cataloguing tool. This tool allows museums to catalogue their educational resources using an on-line form which then outputs metadata (as dictated by the metadata schema) which can be embedded in the museums' HTML pages. The cataloguing tool allows users to select terms from controlled vocabularies and/or provides suggestions from lists of uncontrolled terms. The resulting metadata is sent to the museums for incorporation into their resources.
GEM's cataloguing tool, "GEMCat" was used as a model. The model was adapted for the needs of the Canadian heritage and educational communities. The adaptations related to issues of maintenance, distribution, and technological limitations. For example, GEMCat is used either on- or off-line by a select group of highly trained users. CHIN's cataloguing tool will be available only on-line, to avoid problems of distribution and maintenance among a large, continually changing group of users. GEMCat is a Java application; CHIN's cataloguing tool is web-based, in order to ensure that it can be used with various Web browsers.
For this first phase of the project, the CHIN project team catalogued resources on behalf of the museums; for subsequent phases of the project, the goal is to have the museums catalogue their own resources using the cataloguing tool and the guidelines provided by CHIN. The gateway will quickly become more comprehensive as museums begin to use the cataloguing tool to create the metadata for their own resources.
It is hoped that as metadata begins to play a more central role in other applications, the CHIN cataloguing tool (and the metadata schema) can be extended so that it can be used for content description in disciplines other than education. For example, it could be used to catalogue cultural tourism resources as well as educational resources.
The Metadata Harvester
Another important part of the Learning with Museums gateway is the metadata harvester. After the museums have used the cataloguing tool to create and embed metatags in their on-line resources, the harvester will be used for automated collection of the metadata. The harvester will periodically crawl identified CHIN member museum Web sites, and retrieve the metatags within the educational resources. The collected metadata will then be stored in a periodically updated database that can be searched by users of Learning with Museums.
In the future, it is hoped that the harvester will be able to interpret and collect metadata created by other tools in addition to the CHIN cataloguing tool.
Learning with Museums provides learners with an efficient and easy-to-use resource for accessing on-line educational Canadian museum content. Users of all types can trust that the material that they are accessing and using comes from a credible, authoritative source - the Canadian museum community.
In addition to the benefit to learners and the educational community, the museum community benefits by having their material easily accessible by the target audience(s). The tools and guidelines, provided by the Learning with Museums project, will provide a model for museums wanting to catalogue their material. As the metadata embedded in museums' resources is based on an international standard, the Dublin Core, it is hoped that Internet search engines will also eventually be able to read the embedded metadata and discover the museums' resources.
For this phase of the project, to ensure that Canadian museum materials are well represented in broader directories, the gateway includes reciprocal hyperlinks with Canada's SchoolNet, the Canadian Studies off-line materials database (will be accessible on-line to give access to records of off-line learning materials produced by the federal government, such as books, videos, CD-ROMs etc) and GEM. Because these projects have used common standards, it may be possible to integrate the projects more closely in the future.
Learning with Museums is an exciting step in the goal to further develop access in the distributed environment. Those interested in finding out more or contributing to Learning with Museums, please contact CHIN by telephone 1 800 520-2446 or by e-mail: firstname.lastname@example.org. Learning with Museums will be available, as of May 18, 2000, on CHIN's Website (http://www.chin.gc.ca).
http://wwwx.netheaven.com/~coopercc/xmlparser/intro.html ñ PERL module for parsing XML
http://www.w3.org/RDF/Implementations/SiRPAC/ Simple RDF Parser & Compiler
http://metadata.net/dstc/ Reggie, The Metadata Editor
http://www.ukoln.ac.uk/metadata/dcdot/ UKOLN DC-dot - a Dublin Core generator.
http://www.pault.com/Xmltube/dtdgen.html ñ DTD Generator from RDF
http://www.thegateway.org/ GEM Gateway page
Guidance on expressing DC in RDF
http://www.imsproject.org/metadata/mdbest01.html - IMS Meta-data Best Practices & Implementation Guide
http://www.w3.org/TR/REC-rdf-syntax/ W3C RDF Model and Syntax Specification
http://purl.org/dc/documents/rec-dces-19990702.htm DC Metadata Element Set, Version 1.1: Reference Description
http://archive.dstc.edu.au/RDU/reports/RDF-Idiot/ Good brief explanation of W3C Specs
Annex A: Project Team members and their contributions to Learning with Museums
Annex B: Metadata Schema for Learning with Museums