For background see bugzilla:
http://bugzilla.mozilla.org/show_bug.cgi?id=123124

Problem:

After the JRE is installed if a user installs an instance of some program that uses the gecko rendering engine, that instance will not know about Java and it will attempt to download another copy from the web.

Netscape is willing to alter their installer so that it can detect if a version of the JRE already exists on the machine and do what needs to be done to associate the instance of gecko with that java.

However, since Netscape can't really know what needs to be done to properly connect the plugin to the browser for ever version of the plugin (including future versions). It would make life a lot easier for them (and the end user) if we provided a mechanism that they could used to install us during their install.

Possible implementation:

We create a command in a well know location within the JRE.

This command, when called, will do what needs to be done to connect that JREs version of the plugin to the browser.

It will know what browser it needs to connect to because that information will be supplied via arguments from the calling entity.

The calling entity (the browser install) will know where to find the JRE by looking in the registry (did I mention this is a windows only problem for now).

In this way we can insulate the browsser from having to now the details of plugin installation. Also, if we wanted to, we could use this to track where the plugin has been installed to and unistall it in case that JRE is uninstalled.