Capturing and Presenting Relevant Information in the EDCProject Status(back to homepage) Our progress thus far in the project has been focused on information gathering and initial design. We have broken the problem into three components
Our goal for the project is now to take this information and present it to Hal Eden and Eric Scharff, and to identify our most fruitful coarse of action. First, a thorough workflow needs to be developed that integrates the three components of the project, second we believe that a prototype of this system could be developed in a reasonable time frame, though scope of such must be considered. Information SourcesAs the system we intend to design should be able to retrieve relevant information from a variety of sources, we started by identifying potential information sources. These are listed below.
Periodicals like the Daily Camera are perhaps the most likely place to find information related to Boulder transportation. It should be noted that many periodicals are available in one form or another online. These digital sources are better targets for automated informational retrieval than paper sources are. Secondary news sites tend to filter through content generated by others and provide links to the original sources. If a secondary site was focused on issues relating to Boulder transportation, it might be a good target for finding solid information. Resources who target professional transportation engineers may be of some use to EDC users. If they can be retrieved with some ease, normal citizens could have access to professional sources. Finally, past EDC sessions could be summarized and provided to current EDC users. Information Retreival
In an ideal world, every document we might be interested in would be already categorized for us by RDF. In such a wonderful world, all we would have to do is go out and look for these documents, bring some back to the EDC and use search criteria against their descriptions. Unfortunately, only a small minority of content providers have taken the effort to classify their documents using RDF. A more likely way of gathering documents will be to set a spidering robot out to search a content site. The bot would have to intelligently locate articles that are likely of interest to EDC users and bring those back to the EDC. It is our beliefe that the EDC should maintain a local database that would be populated by these agents for recall as needed during EDC sessions. This would speed the ability of the EDC to recall resources, and allow the gathering agent more time to process online resources. We should also provide some mechanism for manually inputing resources into the reflection space of the EDC. If somebody types up a really great resource on a typewriter, we should be able to scan that in and view it in the reflection space. Information Presentation
the main focus of this project will be providing relevant content to EDC participants in a timely manner. We believe the reflection space os the most likely cantidate for this information, we also believe that this information would best be provided in a "pull" manner rather than "push"ed to the user (in this area we might explore the results of the project on critics in the EDC). Access could be provided either by placing a piece on the board or by a link in the reflection space. Furthermore, we believe that EDC users should be able to directly browse the EDC's repository of information. Current Systems
Current Design:While it would be relatively straight forward to search a single data feed (for instance the Daily Camera online) for transportion articles and present those in the reflection space, a naive approach would make expansion to multiple data feeds difficult. Our design attempts to address the needs of multiple data sources, difficult categorization problems and presentation issues. Keeping this system modular will create an decent oppurtunity for expansion. Architecture Components
|
