Thanks for doing this. Your method for sharing Compendium is better than mine in several ways:
+ Access is browser based, so it's platform independent: Mark K. and other Mac users won't have any problem with it. My version requires users to have a special version of Compendium installed, and at the moment we only have Windows versions.
+ Performance is much better, since it's really just a fancy version of desktop sharing.
Both of our versions currently are challenged by sharing documents. Compendium running locally allows users to drag-n-drop any document into a map and start linking it, but when remote users try to do this the current mechanisms for file sharing break. So the best way to share documents is to put them up on the web someplace (so they have a URL) and then you can drag-n-drop the URL into Compendium and sharing will work fine.
As background for the whole community, multi-user Compendium uses the MySQL open source database system for storing data. Mark's approach, which I prefer to call "Citrix Compendium", uses a beefy server running Citrix's Metaframe Presentation Server. When someone logs in the system starts up an independent instance of Compendium on that server, connected to MySQL (also on that server), and let's the remote user see and control their Compendium instance using terminal server protocols which are by now highly refined. Remote users of Citrix Compendium sometimes find that it seems faster than Compendium running locally on their own computer! The limit of this architecture is that each concurrent Compendium user is consuming a significant portion of the server's CPU, so at some number of concurrent users (we don't know how many) you'll start to see a performance degradation.
The multi-user architecture that I've been working with, which I call "Remote Compendium", connects the Compendium application, running on the user's computer, to a remote MySQL database running on a server at a ISP (Internet Service Provider). Compendium stores and gets it's node and link data by communicating with MySQL over the Internet. In principle this approach scales to any number of concurrent users, but there is a performance penalty: even the smallest packet of data takes about .1 seconds to go between Compendium and MySQL, and in some cases these small delays add up to long pauses. Opening a map, for example, takes about .3 seconds for each node in the map, so a large map with 50 nodes takes about 17 seconds to appear after you double click on the icon. We developed this version for one of my clients (Southern California Edison) as part of switching them over from QuestMap to Compendium.
What would be really great would be to have a multi-user version of Compendium that was architected to be both high-performance over the Internet and to scale nicely as more users come on line. That's just a simple matter of programming ... and we are working on getting some modest funding (about $200K) to get a robut prototype, hopefully this year.