Those interested in this area should read about related works. One is the NSDL effort involving Andries van Dam.
We want a facility to allow easy, (semi)automatic generation of course (training session, or any similar use) web sites based on collections of objects gleaned from CITIDEL. This is the logical progression of the ability to search for useful resources: since CITIDEL indexes and allows retreival/browsing of these resources, it is in the best position to be able to aggregrate them. Building a course web page, manually, is a common task which follows the search for and discovery of resources. We propose to create this facility for automating that task. The details and issues follow.
We envision some kind of aggregation mechanism, perhaps implemented via server-side database records, or cookies (as in retail sites -- think shopping cart), where a user of CITIDEL can flag or drop objects. Also, there would need to be an interface for removing objects and perhaps other operations on the aggregation (such as ordering them or placing them in a hierarchy).
Once the user is done collecting objects, they proceed to create a web site based on them. This could be initated by a menu option when they are viewing the collection (e.g., "checkout" in online shopping). We could make a set of templates available which they can choose from to quickly determine the site structure. The simplest of these could be linear (a list, ordered by appearance of the objects in the collection). More thought needs to go into this.
Once the user is done with the template selection and other customization, the site must be automatically created. An issue here is transfer. Should we archive it and send it as a file? Maybe this is even more work than most people would be willing to do (downloading the archive, putting it on a machine with a web server, uncompressing it). Perhaps we could build the site and then host it on our machines, owned by the registered user/teacher who built the collection.
In the event we want to actually allow the user to host the collection off CITIDEL, there is the issue of linking versus mirroring. Perhaps part of the object metadata could describe whether an object has grants for full mirroring, in which case the CITIDEL user can select to fully retreive it for inclusion in their site.
An important issue related to the usefulness of this project is the metadata format. We need some realistic idea of how objects will be represented on CITIDEL. While it is clear that metadata will lean toward using one or more of the IEEE LOMS, IMS, and DC education standards (and cross-walking between those), it is not clear how far authors can be relied upon to provide rich metadata for their resources using these schemes. Given a reasonable idea of this now, we can create the automatic web site generation system and later plug it into CITIDEL as a module. Even better, given standardization efforts across NSDL, leading to agreement to use the above mentioned formats, our system would be pluggable into a variety of DLs in the NSDL sphere and beyond.
While the normal mode of use in CITIDEL, and the focus for current implementation, will involve users downloading resources and locally building up courses from them, this project will include some brainstorming and summarization of issues like those listed below.