February 20th - March 13th

  Java Team - Kenneth Philbrick

Tasks To Complete:

Kenneth Philbrick
  • Continue to work on the Messageing system
  • Doing UI Work -- For both sides of the project -- Java & one.world -- All Done


Current Status:

Time Spent
Wrote Directory Diff Code. Decided that it was a bad approach
4 hours
Did a little UI work.
2 hours
Did a tone of UI work, I will have the UI released late in the evening on the 27th. Finished up the properties.. All that needs to be done is to load/save the settings. And work in a few necessary interfaces.
9 hours
More UI codeing
8 hours
totally finished UI work, posted final version of ui, did a little bit of preformance monterating of UI.
5 hours
worked on paper
5 hours
Integrated Components, UI, Master Browser
3 hours
Started To test Master Browser local search on large data sets found two small bugs..
3 Hours
Fixed\Re-wrote Jimmy's string parser, added the ability to search against ID3 Tags, Fixed internal data store to hold, more file machine info. Things like files ID3, machines bandwidth, files header, length, and alder32. Also added local search propgation, and local msterbrowser creation.. Installed windows version of VM-Ware with windows me so I can easily begin testing of networking code on my dorm computer.
5 hours
Got my UDP, broadcasting system to work, now UDP, and TCP packets can be used interchangeably.. Started Working on the Login System, broadcast login works, broadcast find parent login works, now working on worst case login synero.. Need to thread login.. Will do this later..
12 hours


Week Summary

Kenneth Philbrick
  • 2/20/2001 So today I wrote the directory scaning code.  The idea was that it would scan a given directory and send out differences / additions / removals. Well I got the code to work, in around 1 hour, but then I spent 3 hours teweeking it and trying to get the code to both be fast and use les memory. The code was a real memory hog. I have decieded that at least in our first release we shouldn't use it and we should just take a snap shot of the directories when quero is started.  I think with all the other places quero is using ram, the directory diff code is just pointless wast.

  • 2/25/2001 nothing much to report.....

  • 2/26/2001 Finished all the other main screens / tabs in quero, wrote the settings window code, I promis the UI code will be released late in the evening on the 27th.. A note for anyone who is working at home.. Download the JMF you will need it to run part of the UI.

  • 3/04/2001 I guess its been a little while since I dumped out a major update.. So here goes.. Well after getting through that paper non-since, and releaseing a version of my ui, I have gone onto integreate all componets and add networking code.. The the one component that is left to be integrated is daves file transfer code.. This will be done after we login, and correctly search. So I found out that sockets, cannot send broadcast packets only datagrams, no big supprise here, but that ment I needed to add a second networking connection point.. Now all Java Quero hosts have a TCP port and a UDP port. The udp port is just used for sending/recieveing broadcast packets. I also found that broadcast packets are reflected back to the sender, which has turned out to be a nice feature for debugging. But packets are now being ignored if they are recieved by the sender.. The basic handshakeing of determing a host to connect to via broadcasting works now, and the onlything left to do on login, is to now write the code which climbs the tree hiearchy, looking for a host..
  • I realized two things.. First there is sorta a bug, and a featuer to the quero system.. The bug is that if hosts try to connect to the quero system via the ping and climb method. The system will result in a fragmented network, but it won't if all hosts enter the system through a single access point.