Status Report: one.world (02/13/2001)

We fell a bit short of the goals we set out for last week. We will need a few more days to meet the goals we set.

Last Week

This Week

Deliverables for next tuesday

By next week, we want a demoable system without any frills.

Time spent

Daniel

Albert

Comments

Daniel

Just some stuff for writing the report later on. If anybody likes to comment on them, I'll be glad to listen.

  1. Chaining events. It would seem useful to have some framework to do within the system as it is fairly common. I managed to hack it right now. The main difficulty is that right now when you have events coming into the EventHandler, you have to 'forward' it to the appropriate component. However this assumes that you know what events the forwarded handler can accept. This situation occured because I needed to use the RemoteManager and my component at the same time which both need to get EnvironmentEvents.
  2. REP is sweet, once you get over the initial learning curve, It would be nice to have some starter code for it since it appears setting it up would be the same for most applications. Enter RemoteMgrWrapper which is a ripoff of the RemoteManagerWrapper that is in TestREP.java. Hopefully this will make its functionality easier to setup.
  3. When two environments are communicating via REP, is the communication bidirectional without one side (client) setting up a named resource?.
  4. Checkpointing. I can't get it to work. I made transisent but I still get the not serializable error. Why can't Java tell me exactly which field is causing me problem. Will linkforcing and making Inbox single threaded solve this problem.

Albert

Two things this week

  1. Swing isn't listed in the protection domain. I'm not sure what this means. Can I use it? The counter does, but it isn't explicitly allowed...
  2. Why was skeletor declared final? It would seem useful to be able to extend the thing to create skeletons for other projects
  3. I'm finding some of the comments in the documentation to be vague. For instance in SimpleOutputRequest, the comments for the constructor parameters are something like, "closure -- The closure for the new simple output request" is a fairly obvious statement and doesn't provide information. It may have been nicer to have no comments than a reflexive one.

Daniel Cheah
Last modified: Tue Feb 13 14:34:49 PST 2001