I've been recoding various routines in DCOMP because the code therein is somewhat awkward and won't work when I add functionality to the system. Some of the code in DCOMP was a hackjob designed to be replaced later after several pieces were finally laid in stone. Thus the reason for my recoding them now.
I'm working on getting group functionality working fully this week. This will entail finishing off the group structure, allowing users to switch into groups, and build their own groups.
Also working on initialization files, which will allow module registration and such. Probably need to do initialization file and processing of that file before I can finish off groups, modules, and applets.... mmmm....
- rebuilt several of the routines in interface to allow for more information about groups from the server
- rebuilt the group structure in the server to include three types of groups:
- Limbo - groups with no leader, no applications. this is the group everyone gets dumped in when they log into the system.
- Default - groups with no leader, aren't deleted from the system when they are empty!, and are created when the server is started as specified in the .adminrc file
- Other - volatile groups with applications and leaders
- almost at the point where people can log in, and switch to the group of their choice.
- decided that modules will log in, send information about their applets (i.e. the html code to embed in the web-page), and then request the downloadgroupinfo if they need such information.
- build a .adminrc (initialization file) and i'm working on the algorithm to process that file correctly and build the necessary structures so the server can use the information.
I talked to Signorile on Friday and we discussed changing the system. I will spend next week looking into RMI, JDBC, and other Java 1.1 features and possibly redesigning the system accordingly.