comPADRE Technology

In the past 18 months, the comPADRE Technology Division has developed a number of applications, code modules, and relational database structures that we feel are worthy of mention. This appendix details what we feel to be some of our most notable achievements.

The foundation of comPADRE and its associated collections is a complex relational database structure housed at the American Association of Physics Teachers. This database was built with many goals in mind: provide content to the physics and astronomy communities in a dynamic and meaningful manner, store item metadata in a Learning Object Model (and DublinCore) compliant fashion, and adhere to the Open Archives Initiative protocols to facilitate metadata harvesting by the NSDL. We have already passed the most stringent automated NSDL analysis available, and are currently working with the NSDL folks to implement the harvesting procedures.

Above this database resides our application and web framework, built with ColdFusionMX and Microsoft SQL server. We have engineered a three-tier environment, encompassing development, staging and production levels. This allows not only for faster and more efficient application and website development, but also provides many layers of security, redundancy, and multi-leveled user interaction.

The heart of all the comPADRE collections is our infrastructure. Built from August - October of 2003, these globally implemented components govern approximately half of our overall functions, including searching, item creation and management, user account administration, user preferences, site security, item relations and annotations, and much more. This infrastructure has a number of purposes; not only does it allow us to test and roll out global code updates in a unified manner, but it also drastically reduces the amount of time and resources necessary to develop and deploy new collections.

The infrastructure also encompasses a number of other functionalities, including:

• Search: using the simple search interface, users are able to comb our database for resources on a number of axes, including subject, author, URL, resource type, and description. An advanced search is also available, which provides users with the capability to filter their results with a greater degree of control. We are currently developing and rolling out a federated search module, which extends our search capabilities outside the boundaries of our data, collecting additional applicable resources from partner sites.
• Filing Cabinet: users are able to 'bookmark' items and arrange them in a flexible folder structure, thereby creating their own, smaller collections. We plan to allow users the option of making their filing cabinets globally accessible to all logged-in users in our next round of development.
• Discussion Forums: Our message board system has proven to be an excellent community-building tool. Users, from the student collection especially, have shown a propensity to post thoughts, questions and opinions about items inside the collection, as well as general thoughts on physics, astronomy and other topics.
• User Profile and Preferences: We allow users a large degree of control over their comPADRE experience, including how much personal information to submit and display, avatar icons and pictures applicable to the discussion forums, as well as personal defaults for ordering, quantity, and filtering of search results.
• Relations and Annotations: Our editors have the ability to inter-relate items in their collection by creating relations and annotations for the items - this metadata is displayed along with the item details, allowing the user to easily view information the editor feels will be of specific value.
• Comments: Users can comment on items, with the comments being displayable to other users across all collections.
• Resource Submission: One benefit of building a community is the breadth of experience and knowledge that the community members bring to the table. By allowing users to suggest quality resources, we can greatly increase our library of high quality online resources.
• Features: Each collection is able to 'feature' certain items - news, special items, images, etc. These items are the best available, those the editor feels will be of the greatest value to their collection's community. The items can be queued by date and will auto-display and deactivate according to parameters set by the editor. Featured items are also more heavily weighted in the return of search results, so that users are exposed to the highest quality content more often.
• Learning Object Creation and Management: The process of initially creating an item, assembling the associated metadata, proofing it, designating collection specific information, and publishing it on the web is what we refer to simply as 'workflow'. It lies at the core of all the comPADRE collections and remains the cornerstone of our development. Editors are able to track items all throughout the creation process, assign the items for metatagging to specific individual accounts, tag items themselves, request that the item be reviewed by qualified individuals, create collection-specific information related to the item, and approve the item for a production setting. We automated as many management tasks as possible in order to lessen the burden of item management.

During the development of the comPADRE collections, our team has dealt with a number of compliance and compatibility issues. Our collections are all fully functional on both the PC and Mac platforms, and can be used through Microsoft Explorer 5+, Netscape Navigator 4+, Safari, and Mozilla browsers. We have worked hard to ensure that our pages are also compliant with Section 508, as well as being properly formatted for XML generation. All of our collections run off CSS, and we plan to author a more centralized style sheet overhaul/upgrade sometime within the next three to six months.

Evaluation and user feedback are both of utmost importance to the success of our development efforts, so we have implemented a number of tools to measure and record overall and specific response to our web presence. In a general sense, we have two applications to track overall usage. The first is WebTrends, a piece of third party software that logs usage and provides us with the ability to generate customized traffic reports. The second application performs a somewhat similar function, but in a more specific sense, as we wrote ourselves specifically for the comPADRE network. In combination, these pieces of software provide us with a high level of granularity in the processing of usage patterns.

We have created and deployed a global survey module, which allows editors to create and post on-the-fly surveys on their websites. The module also gathers statistics and provides detailed response breakdowns. The survey component is currently being used on the Physical Sciences Resource Center to drive part of our assessment effort. We have posted a survey to help us better understand user motivations and experiences with the website. Analysis of this data will help to mold our next round of development, scheduled for this coming summer.

In the interest of furthering community creation, as well as adding additional value to our resources, we are in the process of implementing a multi-leveled review system. Reviews will be of two types - 'expert' reviews, which will be solicited and controlled by the collection editors, and 'peer' reviews, which will be available through the websites and displayable to all users. Items that have been reviewed in the first manner will be flagged as 'reviewed', and will be displayed for users at the top of all search results.