What would be the best approach in using the current API for call monitoring with dynamic CME provisioning (cti watch) to allow for monitoring of all lines.
 
My intial thoughts where something like this:
 
1. Use the HttpXML API to query all extensions periodically OR the provider API to get only active calls (getCalls())
2. For each extension find out all the associated terminals
3. Create address-terminal pairs for all combinations (Virtual type?)
4. For each pair, instantiate them via the API
5. The CME will return the capabilties of each {
 
If cti watch is not present on the CME, montioring is impossible, How can we force it? I saw some mention of:
 
ISExtension a;
a.setISAllowWatch("true");
 
but how to push this config, execCLI(<List>)?
}

Next. assuming cti is active,
 
6. The application can receive events and control calls based on the view of the extensions as where queried at the moment of 1.
 
Is there a more efficent approach?
 
Namely, I would like todo something like:
 
1. Setup provider
2. Force all lines to be watchable
3. Register my CSTA listener(s) for any devices/extensions that come and go from the CME
4. Receive events for all calls. Based on events instantiate address/terminal obejcts to allow for call control.
 
It would be nice to be able to build a map of controllable entities via an event driven model versus requireing a pre-determined view of devices via a query-type approach.

Yes Ive seen everything. However, these do not address my question of clear end-to-end call monitoring approach

Again, I am interested in monitoring all lines registered to a CME without prior knowledge of the configuration infrastructure. Think of it as passive topology learning. The steps I outlined are basic to any call monitoring application. My flow will work but is ineffecient and has some caveats. I am seeking validation of the presence of a better way.
 
If there is no better approach than the one I outlined, then that is all I need as an answer from engineering here.