i-scream documentation viewer

minutes-20010115.txt

Minutes of meeting 15/01/2001 @ 2pm
Location: UKC Computer Science Meeting Room

Present: ab11, ajm4, pjm2, tdb1
Absent: None

The meeting started with a review of the clients. Everyone 
was impressed with the status of them. Work is to continue 
over the next week or so to complete them.

The issue of the server locking up was to be addressed. 
However, in testing during the meeting it was decided that 
it was a problem in the ClientInterface - an issue which is 
still being resolved. Further investigation was put off 
until this has been completed.

Data to be grabbed was discussed. This is becoming an 
important issue as the clients are requiring more precise 
details. It was decided the the "grabbing" part of the host 
should be seperated, for now, from the main program. A Perl 
program will be written which will parse output from various 
system commands (which ports for different OS's) and output 
it in ascii for use by the host application. This will also 
aid porting of the host into C++, which is continuing. It is 
hoped that this "grabbing" code will be moved to C++ 
eventually, but this is a good stop gap, and will also work 
well if moving to C++ cannot be completed.

The group expressed concern on the progress of the host 
application. With work moving fast on the client side a 
solid and reliable host was required. The current host is 
unpredictable and can crash unexpectedly. These issues are 
being addressed. The group also requested that the server be 
"completed" for a release soon, allowing it to be left 
running on Killigrew[1] for host and client testing.

Another issued raised was the client to server protocol. A 
protocol needs to be clearly defined to make server and 
client development easier. AJ & Tim discussed this and will 
produce a spec soon.

Tim raised the issue of Makefiles. The server benefits 
greatly from some extended Makefiles, and the other 
components could use some basic ones to allow quick & easier 
building. This will be looked into.

Finally, it became apparent that the util package of the 
server contains code required in the clients. It was agreed 
that this package should be available seperately in a JAR 
file for use in the clients. Tim agreed to do this, 
arranging Makefiles and changing code to remove dependencies 
with the rest of the server.

The next meeting will be on Monday afternoon at the usual 
time. The meeting room is booked 2pm to 5pm.

[1] - Killigrew is a FreeBSD machine made available by Tim. 
      Continued use is not guaranteed, but it does provide 
      the group with a "shared" user for running the server. 
      This makes stopping/starting/controlling the server 
      possible by anyone in the group - something not 
      available to us on Raptor.