Fifth Meeting

Agenda

1. C++, CVS
2. Next Meeting
3. Vote on Whether People Who Do Specification Should Be Given the Title Technical Lead
4. Evolutionary Development Proposal
5. E-mail List
6. Specification Presentations
7. Discuss the Design / Specification Proposals
8. Consider How to Divide Up the Tasks

Minutes

• Tables in the room were rearranged to form a nice discussion / conference type table
• About nine people showed up

C++, CVS

• Ming has been busy with midterms and hasn't had time to find someone to talk about C++
• Hopefully, can set one up for next week
• Suggestions were given on possible lecturers to talk
• Also, CVS hasn't been set up because of some delay in figuring out how to set it up
• CVS--a system of source control
• All the code source is stored in a central database
• Each developer has their own copy of the files
• When a developer changes the files, they can type a command which will update the database
• Developers can also type another command to load all the changes made by others to their own private copy
• Hopefully things will be set up later in the week

Next Meeting

• Thursday in the same room sounds good

Titles

• People who do specification must spend extra time learning the material and specifying stuff
• Should they be allowed to use some sort of "honorary title" as recognition of this fact?
• Everyone agrees
• Specification people are free to make up a title for themselves (technical lead, member of architecture committee, whatever)

Evolutionary Development

• Basically, this means we'll try to develop things in stages
• Ming presents possible initial stages
• Group discussed it and modified it a bit
• Hopefully each stage will take about a week or two

Stage 1

• Separate renderer with flat-shaded polygons and a static camera
• Colour
• No occlusion
• No meshing
• No parser (only a demo scene)

Stage 2

• Separate renderer with dynamic camera
• Occlusion
• Winged-edge
• Gouraud shading
• Simple polygon parser

Stage 3

• Meshing
• Adaptive subdivision
• Tri-quad tree?

Stage 4

• Integrated renderer?
• Full parser with scene "cleaner?"
• Hierarchical?

E-mail List

• So far, whenever Ming has sent out e-mails, it has been using blind carbon copy so that no one knows who else is on the list
• Should Ming send out the list publicly so that everyone can e-mail everyone else directly?
• Group prefers setting up some group e-mail distribution list
• Ming will look into it, no sure how to go about setting it up

Specification

• Mark Nieweglowski did specification for the renderer
• Specification wasn't quite ready yet (just sketched out somewhat), will present it later
• Mark Visser did specification for everything else
• Highlights of it are here:
• iterators--used a lot and will follow the GNU and Microsoft style
• no matrices (no transforms really needed)
• standard geometry library
• Scene object for holding the meshes (collection of connected faces)
• Polygonal data structure holds polygons
• RadiositySolver with one exposed method Solve() does everything
• Some discussion about efficiency and structure of it
• People who don't understand are invited to talk to us or to e-mail us

Work Break-down

• Basically, the specification will be listed on the web page
• People just look over the list, choose stuff they want to code, and do it
• People should try to e-mail Ming, so that Ming can warn people if there are potential conflicts
• CVS isn't set up yet, but people can still e-mail Ming with any code they've worked on, and Ming will add it to CVS when it's set up
• (Note: people who can't access CVS--likely engineers will probably be able to download all code from the web page, and e-mail Ming with all changes they made)

[Home]