Fourth Meeting

Agenda

1. Next Meeting
2. Begin Design

Minutes

• Meeting was accidentally scheduled for the same time as the Math 235 Linear Algebra midterm
• About seven people showed up, and three more showed up at around 6:30 when the midterm was over
• Meeting was mostly quiet with not much discussion going on
• Afterwards, we had a big discussion about the direction of the group, how it's organized and run, etc.

Next Meeting

• CS241 midterm next week on Thursday (Ming has to attend this one)
• Can get someone else to run the meeting, or can have meeting on Wednesday
• Everyone agrees that Wednesday meeting is ok.

Design

• No one is really clear on the best way to start design
• It seemed that not many people understood radiosity well enough to make suggestions
• People suggest objects that might be useful and can discuss stuff
• Alternately tried a bottom-up and a top-down approach to finding useful objects
• Made a general high level layout of objects
• Below is a cleaned-up diagram of some of the objects we looked at
• (It's not UML, but there should be enough info there to get a feel for the layout of stuff)

Source Control

• I'm personally for manual source control but the group prefers computerised source control
• CVS is available on the undergrad machines
• I'll try to set it up over the weekend

C++

• We'll probably need a "seminar" type thing on this topic
• I'll try to organize one

Other Issues

• Name changed to Student Software Group
• Mark Visser found a free OpenGL compatible layer called Mesa available for most platforms
• Should make programming the final rendering stage of the project easier
• Http link to the appropriate place has been put on the radiosity links page
• Today's meeting was slow, not much participation
• At this rate, project will not get done and people will drop out
• Originally, I envisioned a lot of people participating in design and discussing stuff
• This obviously isn't working
• Project is too complex with not enough people understanding the algorithms to do this
• Could divide project into smaller parts that people could work on
• "Parts" aren't that small though--e.g. the entire geometry library is one part
• All the parts are interconnected so one person should be responsible for coding yet
• But then there won't be enough stuff for people to do
• Also, I don't have a feel for how many people are actually willing to participate in coding etc.
• Timing of the group is also bad (right in the middle of midterms)

Solution

• Part of the problem may the choice of room--the rooms in MC have a classroom feel to them which discourages participation
• Conference rooms in SLC could be booked instead
• Should e-mail everyone to see how many people are still interested so as to get a better feel for how many people are still interested in the project
• Break things into parts which an individual will specify but not actually code
• These specifications can be presented to the group and discussed
• Then work can be divided up
• (This is probably the ideal solution for this particular project--things may be different for other types of projects)

[Home]