Starting a very complex project to use JLine - Advice on best approach

[Joseph Morgan]

I'm at the beginning of a very large project to build an interactive console for managing clusters of servers.  The ultimate goal will have a console server application with a number of global commands, one of which is an ability to switch what I'm calling the "console-mode" to one of three values.   Each of those modes will have significantly different syntax.

The console is actually acting as a single aggregate interface for any given set of connected servers, which can dynamically change as the user chooses.  If the user chooses to act upon 10 servers, when a given command is entered, it is processed and sent to all 10 servers.  Then the user might select to act upon a different set of servers with a different command.

 The number of server configuration commands to date (beyond the global commands)  involves 1555 "objects" having 22821 properties.   All of this needs to be driven by meta-data drawn from instrumentation within the servers themselves.  I know how to extract that meta-data.  

Therefore, aside from the global commands (which will have its own meta-data), all the completers, highlighters, parsers, commands and such must be 1) driven by the meta-data, and 2) be dynamic.   That is, I would love not to have over 10MB of meta-data persistently in memory if it can be avoided.

So far, I'm either finding stupidly simple examples that would lead me to considerable hard-coding, or super-complex examples  (e.g., Karaf) I'm not sure how to properly follow.

I'd love to hear recommendations on the "best" approach, so I start off on the right path.  I'm not new to Java, but am completely new to JLine.  Though I hard-coded a simple example, I'm miles away from my project goal.

All assistance is appreciated.

Thanks.