<snip/>
Actually, I was thinking about a single JVM: merlin would start up via a service and load most
of the "operating system services" it needs; when the user actually wants to interact with the
application, a second container would launch in the same JVM and pop up a gui for
user interaction... (afraid that running in two separate JVMs and remoting all calls between
gui and back-end is going to make the app too slow; but starting everything at once makes
the startup rather slow...). Not feasible?
OK - I now understood your objective.
One question ....
Are you thinking "component browser" (outside looking in) or componet panels (inside looking out).
Don't completely understand the question, basically I'm dealing with an asset
structure (strongly typed hierarchy of "stuff"); I have a number of navigator
components (tree,...) and a set of contextualizable panels for each item in the
hierarchy (e.g. email panel with all mails that have to do with a specific panel);
the idea is that the backend stuff gets loaded once; the physical user then
requests to perform some operations with the assets, he would basically specify
the name of an app that he wants, this app is a component that would be started,
that component then goes and finds the appropriate navigators, item panels,...
and lays all that out for the user to interact with. When user is done, all the components
that have to do with the gui interaction get disposed().
Seems to go more inside looking out... Let me know if my answer is off course...
Thanks, - Filip
Stephen.
- Filip
--------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
--------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
