i-scream documentation viewer

minutes-20010108.txt

Minutes of meeting 08/01/2001 @ 2pm
Location: UKC Computer Science Meeting Room

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

The meeting began in an enthusiastic manner, with all group 
members eager to get on - not what you'd expect for the 
first Monday of the Millennium!

First, a review of work done, if any, over the Christmas 
break. Paul was first to state that the "historal clients" 
he's been doing (web based) are a lot harder than they first 
seemed, but they are coming along very well. Tim reported 
that he had put a queue into the central point of the 
server, and added a few features (eg. javadoc generation) to 
the makefiles. Ash told the group about his work on the 
host, and plans to restructure it to aid porting to C++.

AJ was pleased to report back that it had been lovely and 
sunny in Teneriefe, in constrast to the snow in the UK, and 
that he'd had some 'ideas'.... the group sighed :)

Initial discussions started with what needed to be done this 
term, and when by. It was agreed that there were probably 10 
weeks at the most to get everything finished, and everyone 
was happy with the notion that it should be completed 
*before* the start of project week. A target of around 5 
weeks was set to complete the majority of the coding, giving 
us the rest for documentation, testing and review. More 
precise goals may be set in weeks to come.

AJ mentioned that a "text based" client could be useful, and 
likened it to the "top" program found on raptor 
(/usr/local/sbin/top). This would display basic information 
in an easily digestable form, and would be quick to load. 
This is something that the customer has requested.

AJ also brought up an idea suggested by Tim a while back, 
the restructering of logging in the server. The idea was 
that the log messages would be sent to a logging unit, as 
they do at present, but that this logging unit would pump 
these messages back into the filter system for delivery to 
clients. This would remove the single logging console 
(although it could still log to a file), and allow any 
client to view the current status of the server.

This brought on the discussion of how these would pass "un-
filtered" through the system, which then led to the similar 
issue with heartbeats. This would be looked into by Tim, 
although no work would be done on the logging setup at this 
stage. More emphasis was placed on making the server 
'complete', which entails fixing the heartbeat problem.

Next it was decided what would be done in the immediate 
future. Paul requested doing the command line client after 
completion of the historical clients, which should be 
completed in around a week. The group agreed on this.

AJ said he had ideas for a GUI client, and would work on 
this. Ash said that a rewrite of the host was required to 
make porting to C++ easier. He also said work on the C++ 
host was to begin soon.

Tim is to complete the functionality in the client interface 
part of the server, and patch up other areas that needed 
fixing (eg. heartbeats). The aim is to "complete" the server 
at this stage.

Finally, it was pointed out that the server needs to be 
brought to a stage where it can be released and set running 
on raptor. This is required so the group members working on 
hosts and clients can test their setup is complete.

The next meeting will be on Wednesday afternoon after the 
meeting with Ian. Meeting room is booked 3pm to 5pm.