At 3:00 PM -0700 5/15/08, Jim Slotta wrote: >sounds good - but this is more than what Rokham had done
Yes -- Scott's working on using JackRabbit to not just store OTrunk content as semi opaque blobs of Activity content with additional meta-data -- he's worked out a way for JR to also keep track of all the internal object references that actually make up the essence of an activity -- at least the essence as it relates to introspecting into how the COW (copy-on-write) learner data objects relate to Activity content. What isn't in OTrunk now but is important to add and will probably be quite possible in JR is to add meta data for many of the pedagogically significant objects WITHIN an OTrunk activity. Thhink about extensions of the Model-Data-Activity schema. There are at least four big questions I have about the suitability of JR for LOOPS/TELS: 1) How will a JR repository handle searching and querying extremely large datasets (these data include authored activity, learner and teacher data overlays)? 2) How can we integrate activities with dynamically generated content and processes with JR? 32) Can the an Object transport protocol (layered on top and used to communicate from a learner session to the external OTrunk/JR repository) effectively handle a large number of semi-realtime object updates. For example (assuming a fast network) can the OTObject transport <=> JR Repository handle 20 semi-simultaneous groups of 5 students -- with each student in the group of 5 using their own computer and sharing a whiteboard. I suspect that their might be problems. 4) Can we build replication on top of the JR repositories? This is a critical need to make this work in schools in a large scale. This may not be important for many research projects but is absolutely critical to us. These are definitely issues that Scott and I will be looking into and we would LOVE it if other groups could help in these areas in whatever way they can -- ex: scenarios, simple load testing, research replication architectures, or just being willing to engage in discussions of these issues. >- as they >have been working to revise it into a more general and powerful >solution for the field. I love to read more about the work on this list. >I think they've basically taken Rokham's jackrabbit CMS stuff and >added a new architecture for objects that expands to include semantic >metadata (still not specified) and user interaction data) I'd especially like to hear about how they're thinking about including User Interaction data. To make it work well it needs to be based on a deep understanding of the existing OTrunk architecture and plans for the future. For example in general we've only been creating our COW object proxies for learner data but OTrunk has support (under-specified) for defining multiple overlays of COW Object proxies for different types of users. So for example an activity session could have a layer for activity data, another for learner data and still another for teacher data. Or a single activity session could have five separate learner overlays whith some object views combining the data and some separating them. --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "SAIL-Dev" group. To post to this group, send email to [email protected] To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/SAIL-Dev?hl=en -~----------~----~----~----~------~----~------~--~---
